Bronstein: Move by Move

by Steve Giddins

Contents
About this Publication
Acknowledgements & Bibliography
Series Foreword
David Bronstein: an Appreciation
1 Bronstein-Evans, Moscow 1955
2 Bronstein-Korchnoi, Moscow 1962
3 Bronstein-Winiwarter, Krems 1967
4 Nunn-Bronstein, Hastings 1975/76
5 Bronstein-Rojahn, Moscow 1956
5 Bronstein-Rojahn, Moscow 1956
6 Bronstein-Tal, Riga 1968
7 Bronstein-Ostrup, Gausdal 1994
8 C.Chandler-Bronstein, Gravesend 1996
9 Tal-Bronstein, Moscow 1966
10 Bronstein-Zamikhovsky, Leningrad 1970
11 Bronstein-Hunt, Maidstone 1994
12 Kaplan-Bronstein, Hastings 1975/76
13 Bronstein-Ljubojevic, Petropolis 1973
14 Bronstein-Ratner, Kiev 1939
15 Bronstein-Levenfish, Moscow 1949
16 Sax-Bronstein, Teesside 1975
17 Bronstein-Browne, Reykjavik 1990
18 Bronstein-Botvinnik, 22nd matchgame, Moscow 1951
19 Bronstein-Korchnoi, Moscow 1964
20 Bronstein-Keene, Teesside 1975
21 Bronstein-Najdorf, Budapest 1950
22 Bronstein-Szabo, Budapest 1950
23 Botvinnik-Bronstein, 17th matchgame, Moscow 1951
24 Bronstein-Keres, Gothenburg 1955
25 Bronstein-Golombek, Moscow 1956
26 Zita-Bronstein, Prague 1946
27 Pachman-Bronstein, Prague 1946
28 Reshevsky-Bronstein, Zurich 1953
29 Pein-Bronstein, London 1997
30 Langeweg-Bronstein, Beverwijk 1963
 Back to Contents Page
Bronstein: Move by Move
David Bronstein is one of the greatest and most loved
chess players of all time. He won numerous major
tournaments and for many years he was one of the
world's strongest grandmasters. In 1951 he came
agonisingly close to winning the World Championship
title, drawing 12-12 against the reigning champion
Mikhail Botvinnik. Bronstein was one of the most
creative geniuses the chess world has ever seen, and
he left a legacy of wonderful games. In this book, FIDE
Master Steve Giddins selects and examines his
favourite Bronstein games, brings light to some games
which were previously unpublished, and shows us how
we can all learn and improve our chess by studying
Bronstein's play.
Move by Move provides an ideal platform to study
chess. By continually challenging the reader to answer
probing questions throughout the book, the Move by
Move format greatly encourages the learning and
practising of vital skills just as much as the traditional
assimilation of knowledge. Carefully selected questions
and answers are designed to keep you actively involved
and answers are designed to keep you actively involved
and allow you to monitor your progress as you learn.
This is an excellent way to improve your chess skills
and knowledge.
· Learn from the games of a chess legend
· Important ideas absorbed by continued practice
· Utilizes an ideal approach to chess study

Steve Giddins is a FIDE Master and a former editor of

British Chess Magazine. He spent a number of years of
his professional life based in Moscow, where he learnt
Russian and acquired an extensive familiarity with
Russian chess literature and the training methods of the
Russian/Soviet chess school. He's the author of several
outstanding books and is well known for his clarity and
no-nonsense advice. He has also translated over 20
books, for various publishers, and has contributed
regularly to chess magazines and websites.

Publisher Information

Bronstein: Move by Move

by Steve Giddins
First published in 2015 by Gloucester Publishers plc
(formerly Everyman Publishers plc)
Northburgh House, 10 Northburgh Street, London
EC1V 0AT.
Copyright © 2015 Steve Giddins
The right of Steve Giddins to be identified as the author
of this work has been asserted in accordance with the
Copyrights, Designs and Patents Act 1988.
All rights reserved. No part of this work may be
reproduced, or transmitted in any form or by any means,
without prior permission of the publisher.

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the
British Library.
All other sales enquiries should be directed to
Everyman Chess, Northburgh House, 10 Northburgh
Street, London EC1V 0AT
tel: 020 7253 7887; fax: 020 7490 3708
email: info@everymanchess.com
website: www.everymanchess.com
Everyman is the registered trade mark of Random
House Inc. and is used in this work under licence from
House Inc. and is used in this work under licence from
Random House Inc.

EVERYMAN CHESS SERIES (formerly Cadogan

Chess)
Chief Advisor: Byron Jacobs
Commissioning Editor: John Emms
Assistant Editor: Richard Palliser

Back to Contents Page

 Back to Contents Page

Acknowledgements
My thanks go, as ever, to John and Byron, for
commissioning the book, and then editing and
typesetting it. I am greatly indebted to Tony Stebbings,
Alan Hanreck and Rosemarie Hannan, who kindly
shared their memories of Bronstein with me, and also
supplied the original newspaper column in which the
score of Game 29 was published. Similarly, Cliff
Chandler supplied the score of his game with Bronstein,
which appears as Game 8. Ray Keene kindly lent me
his copy of the Toran/Eliskases book referred to in the
Bibliography, and he and John Nunn both confirmed
points relating to their own games against Bronstein
(Games 20 and 4 respectively). Finally, Matthew Sadler
educated me about the finer points of the Najdorf line
featured in Game 17.
The customary disclaimer of course applies.

Bibliography

200 Open Games, David Bronstein (B T Batsford 1974)

Best Chess Games 1970-1980, Jonathan Speelman
Best Chess Games 1970-1980, Jonathan Speelman
(Allen & Unwin 1983)
David Bronstein: Chess Improviser, Boris Vainstein
(Pergamon 1983)
David Protiv Goliafa, David Bronstein & Sergey
Voronkov (Ripol Classic 2003)
Mezhdunarodni turnir Grossmeisterov, David Bronstein
(Fizikultura I Sport 1983)
Moi Veliki Predshestvenniki, tom.2, Garry Kasparov
(Ripol Classic 2003)
Schoepfergeist der neuesten Schachrichtung, Roman
Toran (W ten Have Verlag 1962, tr. E.Eliskases)
The Art of Chess Analysis, Jan Timman (RHM Press
1976)
The Sorcerer's Apprentice, David Bronstein & Tom
Fürstenberg (Cadogan 1995)

Also the journals: Shakhmaty v SSSR, New in Chess

Magazine, CHESS and British Chess Magazine.

References to 'the computer' generally mean Fritz 12,

although Stockfish 5.64 has also been used in places.

Back to Contents Page

 Back to Contents Page
David Bronstein: an Appreciation
David Bronstein was a truly remarkable figure. Not only
was he one of the greatest players of all time, but he
was also one of the most loved, especially by the
average player. Everybody is aware of the details of his
life, and especially his tied world championship match
with Botvinnik in 1951, and also his famous book on the
Zurich 1953 Candidates. However, as Bronstein himself
put it in the New in Chess interview referred to below,
"I'm more than a few numbers. I'm not Zurich '53 and
12-12!".
Given how much has been written about Bronstein
already, I do not intend to present any sort of biography
or career record here. But for me personally, Bronstein
has the extra fascination that I did actually know him,
albeit only a little, and I even played him. Furthermore,
in his later life, Bronstein had a pronounced connection
with the area of England where I myself live, namely the
county of Kent. He had many close friends at the
Charlton Chess Club, on the southern edge of London,
and frequently stayed in nearby Sidcup, often for weeks
at a time, with the family of Peter and Rosemarie
Hannan. What follows is therefore in the nature of a
Hannan. What follows is therefore in the nature of a
personal appreciation of Bronstein, drawing on my own
limited acquaintance with him, and some recollections
by Rosemarie and other members of Charlton Chess
Club, who have kindly shared their memories with me.
Bronstein had always been a great traveller, who loved
meeting people, especially chessplayers of all levels.
Naturally, as a Soviet citizen, he had always been
restricted in his ability to travel, but, on the other hand,
as one of the country's leading players since the end of
the Second World War, he had been relatively
privileged, and had enjoyed his fair share of foreign
tournaments, at least up until 1976. This was despite
the fact that his father had spent seven years in the
Gulag, from the late 1930s, which would usually have
meant that Bronstein would have been regarded as
"politically suspect".
Everything changed for him in 1976, after Korchnoi
defected. The two had been quite close, and in 1974,
when Korchnoi had played his first match against
Karpov, Bronstein was almost the only Soviet GM who
had been willing to assist Korchnoi with his preparation.
He recalls that it was his suggestion that Korchnoi
defend the Tarrasch French with 3...c5, taking the IQP.
Despite Karpov's legendary prowess in IQP structures,
he was unable to win a single one of the seven games
he was unable to win a single one of the seven games
in which the line appeared in that match, nor the two
further games in which Korchnoi used it in their next
match at Bagiuo City in 1978.
Unfortunately, when Korchnoi requested political
asylum in Holland after the 1976 IBM tournament,
Bronstein was one of the many Soviet GMs who were
called upon to sign an infamous collective letter,
denouncing Korchnoi. To his enormous credit,
Bronstein refused, he and Gulko being the only active
GMs to do so (Botvinnik also refused, but was already
retired at the time, whilst Karpov published another,
personal letter against Korchnoi). The result was that
Bronstein became a 'neviezdny', someone who was not
allowed to travel abroad (or, at least, not outside the
Communist bloc). It was hard to imagine a worse
punishment for the travel-loving Bronstein, and, whilst
most such punishments lasted only a few years, he
remained in that position until the onset of Glasnost at
the very end of the 1980s.
When he was finally free to travel to the West again,
Bronstein made the most of it. Although by then in his
late 60s, he spent a large part of his life in the West,
staying with many amateur chess friends, such as the
Hannans and the Dutch enthusiast Tom Fürstenberg,
with whom he later co-wrote The Sorcerer's Apprentice.
From the early 1990s, Bronstein would often stay for
weeks at a time with the Hannans, at their home in
Sidcup. Charlton Chess Club member Tony Stebbings
told me in an e-mail, that "On these occasions he would
come along with Peter Hannan to Charlton club nights
and go through games he had played, show opening
ideas, play blitz, recount anecdotes and talk at great
length [...] He sometimes turned out for Charlton in the
London and Kent leagues and National Club
Championship, his opponents got a shock when they
saw who they were about to play [...] Once in the
London League his opponent turned up late, sat down
and played without noting who he faced; after the game
he found out to his obvious shock and amazement to
whom he had just lost."
The present book contains two games (nos. 8 and 29)
played by Bronstein in evening league matches for
Charlton.
Tony's Charlton team-mate Alan Hanreck recalls that he
took his greatest-ever individual scalp, thanks partly to
Bronstein. One of the opening ideas Bronstein
demonstrated at the club one evening was an idea
against the Trompowsky (View in Game Format):
1 d4 Nf6 2 Bg5 Ne4 3 Bf4 and now Bronstein's
suggestion was 3...d6. After 4 f3 Nf6 5 e4 it looks as
suggestion was 3...d6. After 4 f3 Nf6 5 e4 it looks as
though White is a tempo up on the Pirc line 1 e4 d6 2
d4 Nf6 3 f3, as he has gained the move Bf4, but
Bronstein's clever idea was to exploit this with 5...e5! 6
dxe5 Nh5!.

Now 7 Be3 dxe5 8 Qxd8+ Kxd8 leaves Black with a

reasonable game, and 7 Qd2 Nxf4 8 Qxf4 Nd7!
regains the pawn, or else offers excellent dark-square
compensation after 9 exd6?! Bxd6 .
Alan later got the chance to play this novelty against
'Tromp' specialist, GM Julian Hodgson, at the 4NCL in
1996. Hodgson preferred 5 Nc3, but after 5...Nbd7 6 e4
e5 7 Be3 Be7 8 Nge2 c6 Black had a reasonable
Philidor position, and went on to beat his powerful GM
Philidor position, and went on to beat his powerful GM
opponent.
Tony's mention of blitz is a convenient moment to show
you a game, which I did not think sufficiently serious to
include in the main section of the book, but which bears
reproducing. It is a great example of Bronstein's
prowess at 5-minute chess, and was played in Moscow
in 1962. One of the aims of the present book is to bring
to light some unknown Bronstein games, as I know
there are many players out there who collect Bronstein
game scores. This game has not been widely
published, as far as I know; I found it in Bronstein's little
beginners' book, of which he was quite proud, but which
is very quirky and not well-known.

D.Bronstein-K.Muchnik
5-minute game 1962
Ruy Lopez
(View in Game Format)

1 e4 e5 2 Nf3 Nc6 3 Bb5 a6 4 Ba4 Nf6 5 0-0 Be7 6

Re1 b5 7 Bb3 d6 8 c3 0-0 9 h3 Na5 10 Bc2 c5 11 d4
Qc7 12 Nbd2 cxd4 13 cxd4 Bb7 14 Nf1 Rac8 15 Bb1
Rfd8 16 d5 Nc4 17 b3 Nb6 18 Bb2 Nbd7 19 Ne3 Bf8
20 Qd2 Nc5 21 Bd3 Nxd3 22 Qxd3 Nd7 23 Qd2 Nc5
24 Nf5 a5 25 Rac1 b4 26 Qg5 Qd7?
Now Bronstein strikes with a combination that most
players would be proud of in classical chess, let alone a
five-minute game:
27 Bxe5! Nxe4
27...dxe5 28 Nxe5 Qc7 29 Nh6+ Kh8 30 Nexf7+ wins.
28 Nh6+ Kh8 29 Rxe4 Rxc1+
29...dxe5 30 Nxe5 gxh6 31 Qf6+ Bg7 32 Rxc8 Rxc8 33
Nxd7 Bxf6 34 Nxf6 was the best chance.
30 Qxc1 Bxd5 31 Rd4 Bxf3
Alternatively, if 31...dxe5 32 Nxe5.
32 Rxd6! Bxd6 33 Bxg7+ Kxg7 34 Qg5+ Kf8 35
Qg8+ Ke7 36 Qxf7# (1-0)
Qg8+ Ke7 36 Qxf7# (1-0)

Bronstein was always a complex character, I suspect,

but by his later years, he had become a curious mixture
of childlike joy and old-age bitterness. The former was
evident from some of the memories relayed by
Rosemarie Hannan. She described Bronstein as being
"like an extra granddad in our house." Indeed, "he would
come back beaming all over his face with carrier bags
full of gifts, which he would produce like a conjurer out
of a hat, chocolates, champagne etc. Once he even
carried back four garden chairs and a sun lounger from
the High Street. He must have carried them down the
road in relays and arranged them at the bottom of the
garden, waiting for me to notice."
The childlike quality of Bronstein was also something I
noticed in my most personal encounter with him, when I
played him at the British Rapidplay Championship. The
game itself was nothing special, as I was totally
overwhelmed by the chance to play such a great player,
and lost very feebly. However, in the spirit of sharing
unknown Bronstein scores, here is the game, for the
benefit of those collectors of such.

S.Giddins-D.Bronstein
British Rapidplay Championship, Leeds 1991
British Rapidplay Championship, Leeds 1991
Grünfeld Defence
(View in Game Format)

1 d4 Nf6 2 Nf3 g6 3 c4 Bg7 4 g3 0-0 5 Bg2 d5 6 cxd5

Nxd5 7 0-0 Nb6 8 Nc3 Nc6 9 e3 Bd7 10 h3 e5 11 d5
Ne7 12 e4 Nec8 13 Bg5 Qe8 14 Rc1 Nd6 15 Re1 h6
16 Be3 Nbc4 17 Qe2 Nxe3 18 Qxe3 Qe7 19 Qc5
Rfc8 20 Red1 a6 21 Nd2 f5 22 Nc4 Qf6 23 Nxd6
Qxd6 24 Qxd6 cxd6 25 h4 h5 26 Bh3 Bh6 27 Rc2

27...Ba4 28 b3 Bd7 29 f3 Rc7 30 a4 Be3+ 0-1

But what really sticks in my mind was the matter of

scoresheets. Being a rapid game, there was no
obligation to keep score, and I did not do so, instead
obligation to keep score, and I did not do so, instead
reconstructing the game afterwards. But Bronstein, in
this and all his other games, meticulously kept score
throughout the game. Furthermore, when the game
finished, he signed his scoresheet and then asked me
to sign it as well. There was no need for this, as the
result was handed in using a separate result slip, but
Bronstein clearly intended to keep the scoresheet of the
game, for his records. Imagine the scene – here am I, a
rank amateur, having just played David Bronstein, and it
is he who is asking me to autograph his scoresheet! He
did the same with every opponent, and I suspect he
retained all the scoresheets until his dying day.
But sadly, there was also an increasing note of
bitterness in the later Bronstein. All of those who knew
him experienced this, and it comes out especially
strongly in his last major interview, given to Dirk Jan ten
Geuzendam of New in Chess in 2001, and published in
issue 2002/1 of that magazine. I strongly recommend
this piece to anyone interested in Bronstein, as it sheds
a very clear light on how he was in his final years (you
can find the interview reprinted in the anthology New in
Chess: The First 25 Years , which I edited).
Life had dealt Bronstein a tough hand in some ways,
especially with the imprisonment of his father as part of
Stalin's crazy purges. He was only a young teenager at
Stalin's crazy purges. He was only a young teenager at
the time, and one can scarcely even imagine how
traumatic it must have been to see his father carried
away, and then to have to live as the son of "an enemy
of the people". There was also the disappointment of
coming so incredibly close to being world champion,
only to miss out at the very last moment. Bronstein's
great friend Tom Fürstenberg, whom I once met in the
Press Room at the Wijk aan Zee tournament, expressed
the view that the 1951 world championship match had
"ruined Bronstein's life", and there is no doubt that it left
a bitter psychological scar.
Bronstein often expressed rather bitter views about
modern chess and the younger generation of
grandmasters. Tony Stebbings recalled Bronstein
telling his Charlton friends that he had been born at the
wrong time, and that younger generations had far
greater opportunities to travel and earn money. He said
much the same thing in his 2001 NIC interview. Like
many of his generation, Bronstein was hit quite hard by
the collapse of the Soviet Union and the associated
economic meltdown and inflation, which left him with an
almost worthless pension. In the NIC interview, he
bitterly lamented how "in Soviet times, a player like me
would have a pension that was five times that of a
doctor". He was deeply offended by the sight of players
doctor". He was deeply offended by the sight of players
such as Kasparov and Kramnik, competing for
million-dollar prizes in world championship matches – in
the 1950s, when he played his match against Botvinnik,
the prize for winning the world championship was about
150 roubles and a slightly larger apartment. Now
players were earning millions and "They think quite
something of themselves [...] all because they can push
a pawn from e2 to e4", as he said derisively to Dirk Jan.
On the other hand, Bronstein was aware of his
tendency to complain, and tried to downplay it. One of
the first things he said in the NIC interview was: "Don't
portray me as a disgruntled old man", and Dirk-Jan
adds "It's a request that he will repeat more than once
today...". Bronstein went on to admit that he was far
better off than many of his generation, millions of whom
died in the war (he was exempted from military service
because of poor eyesight). "At least I could travel and
see something of the world [...] I could buy a nice tweed
jacket or a fine shirt in England, luxuries that [the
average Soviet citizen] could only dream of".
Just as the first draft of this book was finished, my
friend Gerard Welling, the Dutch IM, told me an
interesting anecdote about Bronstein, which again
sheds light on the latter's vulnerability. Gerard played
Bronstein in a rapid event in Holland in 1996 and the
Bronstein in a rapid event in Holland in 1996 and the
two got talking after the game. Bronstein started telling
of how, a few months earlier, he had returned to the
small village in the Ukraine where he had been born.
He chatted with an old man, who was farming in the
area. The man had clearly never heard of Bronstein,
and asked what he had done with his life. Bronstein
replied that he had become a chess grandmaster and
had tied for the world championship, whereupon the old
man just grunted: "Huh! Nothing worthwhile, then!".
Most people would just have dismissed the old man as
rude, but Bronstein had clearly been deeply hurt by the
remark, and, even months later, he sat at the table with
Gerard, muttering, "He's right, of course. I have wasted
my life, wasted it completely".
Therein lies the essence of the contradiction in David
Bronstein. On the one hand, he was very privileged by
Soviet standards, but on the other, he had to take some
hard blows in his life, and clearly always felt he was on
the outside, as far as the Soviet chess authorities were
concerned. This was a difficult path to tread, and the
disappointments he suffered took their toll on him. But
his essentially childlike qualities and deep love of chess
never failed to shine through, despite the black
moments. And, as a player, he was one of the most
creative geniuses the game of chess has ever seen,
creative geniuses the game of chess has ever seen,
who left behind a legacy of wonderful games. I have
done my best to choose a representative sample here,
and to annotate them as instructively as I can, using
modern silicon help to penetrate as many of the
mysteries as possible. I hope the book brings pleasure
to Bronstein's numerous admirers.
To finish this small personal appreciation, here is
another game, which was not serious enough to
warrant a place in the main games section, but which
always makes me laugh out loud, whenever I see it. It
was played against a much weaker opponent, at the
Norwegian Open at Gausdal in 1990. For me, it typifies
Bronstein's attitude to the game. It would have been all
too easy for him to have beaten such opponents on
sheer technique, especially after such a ludicrous
opening as White played here (with all due respect to
my friend, Mike Basman), but that was not Bronstein's
way. He chooses the most combinative path (2... Bxg4
is supposed to be the one line that gives White chances
in this opening) and wins with elegance. "Chess should
give people pleasure", he said in his NIC interview, and
he lived up to that principle.

B.Olsson-D.Bronstein
Gausdal 1990
Gausdal 1990
Grob Opening
(View in Game Format)

1 g4 d5 2 Bg2 Bxg4 3 c4 c6 4 cxd5 cxd5 5 Qb3 Nf6 6

Qxb7 Nbd7 7 d4 Rb8 8 Qxa7 Qc8 9 Bf4 e5 10 dxe5
Bc5 11 Qa4 Rb4 12 Qc2 Rxf4

13 exf6 Bxf2+ 14 Kd1 Rc4 15 fxg7 Rg8 16 Nc3 Rd4+

17 Kc1 Be3+ 18 Kb1 Rd2 19 Qxh7 Nf6 20 Qh4 d4 21
Qxf6 Rxb2+ 22 Kxb2 Qxc3+ 0-1

It was a privilege to have known and played you, David

Ionovich!

Steve Giddins,
Steve Giddins,
Rochester, Kent,
February 2015

Back to Contents Page

 Bronstein − Move
1 by Move

6
Trompowsky Idea
A45 XHGFEDCBAY
[Steve Giddins] 1RsNLmKQ+NtR!
1.d4 ♘f6 2.♗g5 ♘e4 3.♗f4 Diagram
2zPP+-+PzPP" 2
XHGFEDCBAY 3-+P+-+-+#
1RsNLmKQ+NtR! 4+-vLPzP-+-$
2zPPzPP+PzPP" 1 5-+-+-+-+%
3-+-+-+-+# 6+-sn-zp-+-&
4+-vLnzP-+-$ 7pzppzp-zppzp'
5-+-+-+-+% 8tr-vlkwqlsnr({
6+-+-+-+-& xhgfedcbay
7pzppzppzppzp' it looks as though White is a tempo up on the
Pirc line 1 e4 d6 2 d4 Nf6 3 f3, as he has
8tr-vlkwqlsnr({ gained the move Bf4, but Bronstein's clever
xhgfedcbay idea was to exploit this with 5...e5! 6.dxe5
Diagram
and now Bronstein's suggestion was d6
. After 4.f3 ♘f6 5.e4 Diagram
XHGFEDCBAY
[ Alan later got the chance to play this 1RsNLmKQ+NtR!
novelty against 'Tromp' specialist, GM 2zPP+-+PzPP" 3
Julian Hodgson, at the 4NCL in 1996.
Hodgson preferred 5.♘c3 , but after ♘bd7 3-+P+-+-+#
6.e4 e5 7.♗e3 ♗e7 8.♘ge2 c6
Black had a reasonable Philidor position,
4+-vLP+-+-$
and went on to beat his powerful GM 5-+-zP-+-+%
opponent. ]
6+-sn-zp-+-&
7pzpp+-zppzp'
(Diagram 2)
8tr-vlkwqlsnr({
xhgfedcbay
♘h5! . Now 7.♗e3
[ and 7.♕d2 ♘xf4 8.♕xf4 ♘d7!
regains the pawn, or else offers excellent
dark-square compensation after 9.exd6?!
♗xd6 ]

Bronstein − Move by Move

 Bronstein − Move
2 by Move

7...dxe5 8.♕xd8+ ♔xd8 leaves Black with XHGFEDCBAY

a reasonable game.
1-mK-tR-tR-+![
7 C99
2+PzP-+-vLP" 5
Bronstein,D 3P+N+-+P+#
Muchnik,K
5-minute game 1962
4+-+P+-zp-$
[Steve Giddins] 5-wQNzpPsn-zp%
1.e4 e5 2.♘f3 ♘c6 3.♗b5 a6 4.♗a4 ♘f6 6+-+-zp-+-&
5.0-0 ♗e7 6.♖e1 b5 7.♗b3 d6 8.c3 0-0 7pzpp+q+l+'
9.h3 ♘a5 10.♗c2 c5 11.d4 ♕c7 12.♘bd2
cxd4 13.cxd4 ♗b7 14.♘f1 ♖ac8 15.♗b1 8+kvl-trr+-(
♖fd8 16.d5 ♘c4 17.b3 ♘b6 18.♗b2
♘bd7 19.♘e3 ♗f8 20.♕d2 ♘c5 21.♗d3
xhgfedcbay
♘xd3 22.♕xd3 ♘d7 23.♕d2 ♘c5 24.♘f5 Now Bronstein strikes with a combination
a5 25.♖ac1 b4 26.♕g5 Diagram that most players would be proud of in
XHGFEDCBAY classical chess, let alone a five-minute game:
27.♗xe5! ♘xe4
1-mK-tR-tR-+! [ 27...dxe5 28.♘xe5 ♕c7 29.♘h6+ ♔h8
2+PzP-+-vLP" 4 30.♘exf7+ wins. ]
28.♘h6+ ♔h8 29.♖xe4 ♖xc1+
3P+N+-+P+# [ 29...dxe5 30.♘xe5 gxh6 31.♕f6+ ♗g7
32.♖xc8 ♖xc8 33.♘xd7 ♗xf6 34.♘xf6
4+-+P+-zp-$ was the best chance. ]
5-wQNzpPsn-zp% 30.♕xc1 ♗xd5 31.♖d4 ♗xf3 Diagram
[ Alternatively, if 31...dxe5 32.♘xe5 . ]
6+-+-zp-+-&
7pzpp+-wql+'
(Diagram 6)
8+kvl-trr+-({
xhgfedcbay
♕d7? Diagram

(Diagram 5)

Bronstein − Move by Move

 Bronstein − Move
3 by Move

XHGFEDCBAY 9
Olsson,B
A00

1-mK-+-wQ-+![ Bronstein,D
2+PzP-+-+P" 6 Gausdal 1990
[Steve Giddins]
3P+l+-+P+#
1.g4 d5 2.♗g2 ♗xg4 3.c4 c6 4.cxd5 cxd5
4+-+-tR-zp-$ 5.♕b3 ♘f6 6.♕xb7 ♘bd7 7.d4 ♖b8!
5-+-vL-+-zp% 8.♕xa7 ♕c8 9.♗f4 e5 10.dxe5 ♗c5
11.♕a4 ♖b4 12.♕c2 ♖xf4 13.exf6 ♗xf2+
6sN-+-zp-+-& 14.♔d1 ♖c4 15.fxg7 ♖g8 16.♘c3 ♖d4+
7pzpp+q+-+' 17.♔c1 ♗e3+ 18.♔b1 ♖d2 19.♕xh7 ♘f6
20.♕h4 d4 21.♕xf6 ♖xb2+ 22.♔xb2
8mk-vl-tr-+-( ♕xc3+
xhgfedcbay 0-1

32.♖xd6! ♗xd6 33.♗xg7+ ♔xg7 34.♕g5+

♔f8 35.♕g8+ ♔e7 36.♕xf7# 10 C78
1-0 Bronstein,D
Evans,L
USSR-USA match, Moscow (Game 1) 1955
8 D76 [Steve Giddins]
Giddins,S
Bronstein,D 1.e4 e5 2.♘f3 ♘c6 3.♗b5 a6 4.♗a4 ♘f6
British Rapidplay Championship, Leeds 5.0-0 b5 6.♗b3 Diagram
[Steve Giddins]
(Diagram 7)
1.d4 ♘f6 2.♘f3 g6 3.c4 ♗g7 4.g3 0-0
5.♗g2 d5 6.cxd5 ♘xd5 7.0-0 ♘b6 8.♘c3
♘c6 9.e3 ♗d7 10.h3 e5 11.d5 ♘e7 12.e4
♘ec8 13.♗g5 ♕e8 14.♖c1 ♘d6 15.♖e1 h6
16.♗e3 ♘bc4 17.♕e2 ♘xe3 18.♕xe3 ♕e7
19.♕c5 ♖fc8 20.♖ed1 a6 21.♘d2 f5
22.♘c4 ♕f6 23.♘xd6 ♕xd6 24.♕xd6
cxd6 25.h4 h5 26.♗h3 ♗h6 27.♖c2 ♗a4
28.b3 ♗d7 29.f3 ♖c7 30.a4 ♗e3+
0-1

Bronstein − Move by Move

 Bronstein − Move
4 by Move
XHGFEDCBAY
1-mKR+QvLNtR!
2zPPzP-zPPzPP" 7
3-+N+-+L+#
4+-+P+-+-$
5-+-zp-+p+%
6+-sn-+n+p&
7pzpp+pzp-+'
8tr-vlkwql+r({
xhgfedcbay
d6!? An unusual move order, which has
nothing special to recommend it, but which
can occasionally catch out the unwary.
[ The move 5...b5 is usually the prelude to
the so-called 'New Archangel' characterized
by the move 6...♗c5 , which is nowadays
quite popular at top level and has
accumulated an alarmingly extensive body
of theory. ]
[ That move is itself a variation on the
original Archangel Variation, which was
marked by the move 6...♗b7 . ]
7.c3
[ Q U E S T I O N : H o l d o n ! I s n ' t 7.♘g5
somewhat embarrassing for Black here?
ANSWER: Actually, it is almost a trap, and
is rather what Black is hoping for. He
replies with the forced d5 8.exd5 ♘d4
, when he has good counterplay.
QUESTION: This looks a bit like the Two
Knights!
ANSWER: It is similar, yes, but here,
Black has a much better version. His
queen's knight has been able to occupy a
central square on d4, rather than the
miserable a5-square, which is its usual fate
Bronstein − Move by Move
in the Two Knights. From d4, it can
eliminate the white Spanish bishop, whilst
the pawn sacrificed on d5 will soon fall.
White can win a pawn by 9.d6 ♘xb3
10.dxc7 ♕xc7 11.axb3 , but Black has
excellent compensation, in the shape of his
two bishops and better development. ]
[ Instead of all this, Bronstein settles for
quite development, with a typical Spanish
move, 7.c3 . With his e5-pawn now
defended, Black was threatening 7...Na5,
swapping off the Spanish bishop, so
White secures it a retreat, whilst
preparing the central advance d2-d4. ]
7...♘a5
[ Q U E S T I O N : W h a t a b o u t 7...♗g4
here? That looks very natural.
ANSWER: It does, and it is often played in
this position, but is generally frowned upon.
The point is that, in these Spanish positions,
...Bg4 is generally a good reaction to d2-d4,
but is less good if played before the pawn
has come to d4.
QUESTION: Why?
ANSWER: Because White can react with
8.d3 , setting up a very solid structure, and
then follow up with Re1, Nbd2-f1-g3 and
drive the bishop away. With the white
pawn on d3, the bishop is not really very
effective on g4, because it does not exert
any pressure on the d4-square. We see
this theme in Game 3 (as will become
apparent, to assist the reader I have
arranged the order of games in this book
by opening, and chronologically within
each opening). ]
8.♗c2 c5 9.d4 ♕c7 10.♘bd2 Diagram
(Diagram 8)
 Bronstein − Move
5 by Move

XHGFEDCBAY Usually, in such positions, the bishop

develops to e7, and then later on, after
1-mKR+QvL-tR! castling, Black frequently plays ...Re8, ...Bf8,
2zPPzP-sNLzPP" 8 ...g6 and ...Bg7, transferring the bishop to a
better square, but at the cost of considerable
3-+N+-zP-+# time. Evans therefore reasons that he may as
well play ...g6 and ...Bg7 at once, thus
4+-+PzP-+-$ saving a considerable quantity of tempi.
5-+-zp-zppsn% QUESTION: So what is wrong with that? It
sounds like a good idea.
6+-sn-zp-+p& ANSWER: Well, as I said, I think the idea
7pzpp+-wq-+' itself is excellent, and in the face of routine
play from White, such as Re1, Nf1, etc, it
8tr-vlk+l+r({ would have worked well. However, the
xhgfedcbay drawback of Evans' move is that the black
king remains in the centre, and a move or
g6?! Diagram two away from castling, which Bronstein
[ The usual 10...♗e7 was therefore safer. ] seizes on for an energetic response. 11.b4!
An excellent retort, and a common idea in
XHGFEDCBAY these Chigorin structures, if Black plays too
slowly. White forces lines open, to exploit
1-mKR+QvL-tR![ his superior development.
2zPPzP-sNLzPP" 9 cxb4
[ QUESTION: Shouldn't Black keep lines
3-+N+-zP-+# closed with simply 11...♘c6 ?
ANSWER: He would like to, but
4+-+PzP-+-$ unfortunately, the pressure against his
5-+-zp-zppsn% centre means that he cannot really allow
White to take on c5. Bronstein then gives
6+psn-zp-+p& 12.bxc5 dxc5 13.♘xe5 ♘xe5 14.dxe5
7p+p+-wq-+' ♕xe5 15.♘f3 ♕xc3 ( 15...♕c7
is possibly slightly better, but even here,
8tr-vlk+l+r( after 16.e5 ♘d7 17.♖e1 , the threat of e5-
xhgfedcbay e6 is terrible, and ♘b6 18.a4 creates new
threats against the b5-pawn) 16.♖b1
QUESTION: What is going on here? Don't "with irresistible threats". ]
you like 10...g6 then? 12.cxb4 ♘c6
ANSWER: I like the idea very much, but [ QUESTION: What about 12...♘c4 now?
unfortunately, it is not effective in this ANSWER: Then Bronstein gives 13.♘xc4
position. What Evans is trying to do is reach bxc4 14.♗b2 ♘d7 ( 14...c3 15.♖c1!
an ideal set-up for Black, where he could is the very nasty point, when the c3-pawn
potentially save himself a couple of tempi. is not long for this world) 15.♗a4 ♗g7

Bronstein − Move by Move

 Bronstein − Move
6 by Move
16.♖c1 and the c4-pawn is doomed. ]
13.♗b2 ♗g7
[ QUESTION: Can't Black take the pawn
with 13...♘xb4 , attacking the bishop on
c2?
ANSWER: Yes, but after the simple reply
14.♗b1 , he still has serious problems. The
pawn on e5 is hanging and cannot be
defended in any comfort: for example, ♘c6
15.a4 b4 16.♕c2 , pinning the knight, and
threatening both 17 d5 and 17 dxe5. Black
has no satisfactory reply, and will
inevitably lose back the e5-pawn, with
his kingside still undeveloped and his
king exposed in the centre. ]
Evans tries to avoid this by hastening to
remove his king from the centre, but his
problems remain.
14.♖c1 ♗b7 15.♗b3 Diagram
XHGFEDCBAY
1-mKR+QtR-+!
2zPPzP-sN-vLP" 10
3-+N+-+L+#
4+-+PzP-zP-$
5-+-zp-+p+%
6+psn-zpn+p&
7pvlp+-wql+'
8tr-+k+-+r({
xhgfedcbay
♕e7 Diagram
[ 15...♕b8 is the alternative, but then White
h a s t h e c o m b i n a t i o n 16.dxe5 ♘xe5
( 16...dxe5 17.♖xc6 ♗xc6 18.♘xe5
is even worse than the game) 17.♗xe5
dxe5 18.♗xf7+ ♔xf7 19.♘g5+ ♔e7
20.♕b3 ♕g8 21.♖c7+ ♘d7 22.♕d3
Bronstein − Move by Move
, with a winning attack. Yet again, the
black king is caught in the centre, with
devastating effect. ]
[ EXERCISE: How can White prevent the
escape of the black king after 15...♕e7 ? ]
XHGFEDCBAY
1-mKR+QtR-+![
2zPPzP-sN-vLP" 11
3-+N+-+L+#
4+-+PzP-zP-$
5-+-zp-+p+%
6+psn-zpn+p&
7pvlpwq-+l+'
8tr-+k+-+r(
xhgfedcbay
It looks as though Black has managed to
shore things up, and is now ready to castle.
ANSWER: 16.♖xc6! This exchange sacrifice
cries out to be played, but, as we will see,
Bronstein had to calculate quite a long way,
in order to prove that it works.
♗xc6 17.dxe5 ♘h5 Diagram
[ The first point is that 17...dxe5? 18.♘xe5
attacks both c6 and f7, and is obviously a
disaster for Black. His 17th move is
th eref ore his only hope, and, indeed,
Bronstein even goes so far as to suggest
that Evans may have seen this far and relied
on 17...Nh5, when playing his 10th move. I
am not so sure about that, but, in any event,
Bronstein certainly had to consider the
consequences of this move, before
sacrificing the exchange. ]
 Bronstein − Move
7 by Move

XHGFEDCBAY ( 21...♘f4+ 22.♔h1 ♕xd6 23.♘xf7

wins ) 22.♗xe5 f6 23.♗g3 ♘g5 24.h4
1-mKR+Q+-+![ ♘f7 25.♕a1 White gains an important
2zPPzP-sN-vLP" 12 tempo over the game. ]
21...♗xe5 22.♗xe5 f6 23.♗g3 ♔f8 24.♔g2
3-+N+-+L+# Interestingly, Bronstein's earlier notes were
also critical of this move, but the computer
4+-+P+-zP-$ thinks it is in fact best.
5n+-zP-+p+% [ B r o n s t e i n r e c o m m e n d e d 24.♗d5
, aiming to exchange off Black's strong
6+p+-zpl+p& bishop and set up the knight transfer Nd2-
7pvlpwq-+-+' b 3 - c 5 , b u t a f t e r ♗xd5 25.exd5 h5
Black's position is better than in the game,
8tr-+k+-+r( though still pretty awful. ]
xhgfedcbay 24...♘g5 25.h4 ♘f7 Diagram
[ A N S W E R : A f t e r 25...♘xe4
18.g4! White continues 26.♘xe4 ♗xe4+ 27.f3
[ 18.exd6 ♕xd6 19.♗xg7 ♘xg7 ♗c6 and now the computer's 28.♖e1 ( or
does not achieve anything, so the knight 28.g5; are both even stronger than
must be driven away from the defence of Bronstein's suggested 28.♕d4 , although
the bishop on g7. ] White's advantage is decisive in all cases. )]
18...♘f4 19.exd6 ♕d7
[ Q U E S T I O N : W h a t a b o u t 19...♕f8
, which simply defends the g7-bishop?
XHGFEDCBAY
ANSWER: Then White just continues 1-+R+Q+-+![
20.♘e5 and if the bishop moves away, such 2+KzP-sN-+P" 13
as with ♗d7 , then just 21.♘xd7 ♔xd7
22.e5 gives White an overwhelming 3-vL-+-+L+#
position. He has three pawns for the
exchange, and his two bishops and the
4zPP+P+-zP-$
pawns on d6 and e5 are slaughtering the 5-+-+-+p+%
black position. ]
Evans' move attacks the g4-pawn, which adds 6+pzp-zPl+p&
a further complication to the mix, but 7p+n+q+-+'
Bronstein had seen everything.
20.♘e5 ♘h3+ 21.♔h1 8tr-mk-+-+r(
[ Bronstein's annotations to this game in
'The Sorcerer's Apprentice' are quite brief,
xhgfedcbay
and he does not comment on this move, but QUESTION: Isn't the pawn on e4 hanging?
in his earlier notes in 'Shakhmaty v SSSR, 26.♕a1! An extremely unpleasant move,
9 / 1 9 5 5 ' , h e p o i n t e d o u t t h a t 21.♔g2 which is a recurring theme.
is slightly more accurate. Then after ♗xe5 ♘e5

Bronstein − Move by Move

 Bronstein − Move
8 by Move
[ Black has no good way to protect the f6-
pawn, since 26...♔g7 is met by 27.g5 . ]
27.g5! Even better than taking on e5.
QUESTION: Reti would have approved!
ANSWER: Yes, the Hypermoderns would all
have been delighted with the way the white
queen operates from the corner of the board,
and t h e g- pawn un der mi n es t h e en emy
fortifications on the long black diagonal.
♖e8
[ Bronstein's criticism of his earlier play
was based on the claim that here, Black
could have made White's task far from
s i m p l e w i t h 27...♕xd6 28.gxf6 ♖e8
. However, the computer shows that this
does not really help, and White crashes
t h r o u g h a f t e r 29.f7! ♖e7 30.♖c1!
. The threats of Rxc6 and Rc5 are
unstoppable, and Black is tied hand and
foot. ( Instead, Bronstein only considered
t h e i n f e r i o r 30.♗xe5 ♕xe5 31.♕xe5
♖xe5 32.♖c1 ♗b7 33.f3 ♖e7
, which is indeed still not fully clear. )]
28.♕d4
[ 28.gxf6 ♘f7 29.♕d4 was equally good. ]
28...♕g4 29.f3 ♕d7 30.gxf6 ♘f7 31.e5
The position is picturesque in the extreme.
The white pawns are advancing down the
board, like Space Invaders, crushing all
befo re th em. The g ame p r es en ts a nice
parallel with Game 5 against Rojahn.
g5
[ 31...♘d8 attempts to hold up the e-pawn,
but then, for example, 32.♘e4 followed by
Nc5 or Ng5 is crushing. ]
32.♖e1 gxh4 33.e6 ♕c8 34.♕xh4
[ The human move, but the computer points
out that 34.d7 is actually mate in 8,
improbable as this may seem. ]
[ Bronstein also points out that 34.♕g4
"would have forced immediate resignation",
but both players were in serious time-
Bronstein − Move by Move
trouble at this point, and so the rest
happened at great speed. In any event,
White's position is overwhelming and he
duly wraps up the full point. ]
34...♖xe6 35.♖xe6 ♗e8 36.♖e7 ♕c1
37.♖xe8+ ♔xe8 38.♕e4+ ♔d8 39.♕e7+
♔c8 40.♗e6+
[ 40.d7+ mates a couple of moves quicker,
but who is really counting at this point? ]
40...♔b8 41.♕c7+
[ Here too, 41.d7+ mates in a few moves. ]
41...♕xc7 42.dxc7+ ♔b7 43.♗xf7 ♖f8
44.♗d5+
1-0
11 C83
Bronstein,D
Korchnoi,V
Leningrad-Moscow m., Leningrad (Game 2)
[Steve Giddins]
1.e4 e5 2.♘f3 ♘c6 3.♗b5 a6 4.♗a4 ♘f6
5.0-0 ♘xe4 The Open Spanish, which has
always been a favourite of Korchnoi. He was
introduced to it by the legendary Leningrad
trainer, Vladimir Zak, who was always a
great lover of the line.
6.d4 b5
[ Q U E S T I O N : T h e m o v e 6...exd4
loses by force, doesn't it?
ANSWER: Actually, no. There are
considerable complications after 7.♖e1
, but it is not clear that Black is so badly
off, strangely. The main line is d5 8.♘xd4
♗d6 9.♘xc6 ♗xh2+ 10.♔h1! ♕h4
11.♖xe4+! dxe4 12.♕d8+ ♕xd8
13.♘xd8+ ♔xd8 14.♔xh2 , with an
endgame where White has two pieces for
rook and two pawns. It has always been
regarded as better for White, but this is
not 100% clear. Unfortunately, though,
 Bronstein − Move
9 by Move
Black's main problem is that he is only
likely to draw at best, which makes it less
attractive. ]
7.♗b3 d5 8.dxe5 ♗e6 9.c3
[ 9.♕e2 , ]
[ 9.♘bd2 , ]
[ and 9.♗e3 are all possible, ]
[ but 9.c3 has always been the main line.
QUESTION: And what essentially is going
on here? Black's position looks quite
attractive – he has free play, a strong
knight on e4 and a queenside pawn
majority. Why shouldn't he have a great
position?
ANSWER: Dr. Tarrasch would be proud of
you! The great German player always
argued just as you do, claiming this was
the only correct defence to the Spanish,
but most great masters have preferred
White here.
QUESTION: Why?
ANSWER: The problem is that, when one
looks a little more deeply, things are not
quite so wonderful for Black. Firstly, his
knight looks great on e4, but is not
entirely stable – in some lines, White can
play c4, exploiting the overloading of the
d5-pawn, which must defend e4, and in
other variations, White can kick the
knight with f3. Secondly, the white pawn
on e5 means that Black must always be
careful to ensure that his king does not
come under attack. And playing ...f5 or ...
f6 himself weakens the kingside further,
as well as the diagonal a2-g8. Finally, the
black queenside pawns have advanced
rather prematurely, and there is a danger
that the c-pawn will end up backward.
Capablanca, in 'Chess Fundamentals', gives
an example of a nightmare scenario for
Black, where White plays Nd4, forces the
exchange, recaptures cxd4, and then
Bronstein − Move by Move
crushes Black on the dark squares,
exploiting the hole on c5. Such a scenario
is only possible if Black mishandles it, but
it is a frequent threat. ]
9...♗e7
[ This old line is somewhat discredited
these days, with the more active moves
9...♘c5 ]
[ and 9...♗c5 both enjoying a better
reputation. ]
10.♗c2 0-0 11.♕e2
[ 11.♘bd2! is more precise. ]
11...f5
[ In his notes in 'David against Goliath',
Bronstein points out that the players had
reached the same position just two days
earlier, at the Soviet Team Championships.
T h e r e K o r c h n o i h a d c h o s e n 11...♘c5
and the game had been drawn. ]
12.exf6 ♗xf6? This pawn sacrifice was
obviously Korchnoi's idea, but it proves
insufficient.
[ Instead, Black could equalize with
12...♖xf6 : for example, 13.♘g5 ♗c5!
"with very interesting complications"
(Bronstein). The computer supports this
view, one variation running 14.♘xe4 dxe4
15.♕xe4 ( 15.♗xe4? ♗c4 ) 15...♗f5
16.♗b3+ ♔h8 17.♕h4 ♗xb1 18.♖xb1
♖xf2! 19.♕xd8+ ♖xd8 20.♖xf2 ♖f8
and Black regains his pawn, with equal
chances. ]
13.♘bd2! This is the point. White exploits
the pin on the undefended bishop at e6, to
win a pawn.
[ Instead, 13.♗xe4? dxe4 14.♕xe4 ♗d5
would be much better for Black. ]
13...♗f5
[ 13...♘c5 is met by 14.b4 ♘xb4 15.cxb4
♗xa1 16.bxc5 with a clear advantage. The
two knights would be much stronger than
the rook in this position.
 Bronstein − Move
10 by Move
QUESTION: Why is that? Black has rook
and pawn against two minor pieces, which
is at least mathematical equality.
ANSWER: Yes, but in the middlegame,
there are usually insufficient open lines
for th e r oo ks to ex ert th eir maximu m
strength. Rooks tend to become much
more effective in the endgame, when the
board is relatively empty and there are
numerous open files for the rooks to use. In
a middlegame such as this, they find it
mu ch harder to develop full activ ity,
wher eas the wh ite min o r pieces hav e
numerous strong outposts, such as d4, e5,
etc. ]
14.♘xe4 ♗xe4 15.♗xe4 dxe4 16.♕xe4
♕d7 QUESTION: I don't really understand
what is going on here. Why isn't Black just a
pawn down for nothing?
ANSWER: Well, it is not quite nothing, but
it is not enough.
QUESTION: So why did Korchnoi play like
this?
ANSWER: I think he just underestimated
White's chances. Black has some
compensation, in the form of better
development. All of his pieces are very active,
and he will gain a further tempo by bringing
his last undeveloped piece, the rook from a8,
to e8, with tempo. Korchnoi simply
overrated his compensation, I think.
17.♗f4!? QUESTION: Isn't the bishop
tactically vulnerable on this square?
ANSWER: Yes, it is a bit, and I am not
convinced it is the best move, although
Bronstein awards himself an exclamation
mark. He admits that the move is risky, but
makes the typically paradoxical comment "I
know from experience that risky moves are
often the most logical".
QUESTION: What does that mean?
ANSWER: Frankly, I don't know.
Bronstein − Move by Move
[ I n a n y e v e n t , I p r e f e r 17.♗e3
, which Bronstein rejected as passive, ]
[ but perhaps the most logical move of all is
to postpone the decision and first move the
queen away with 17.♕c2 , since she will
have to move anyway, and then decide
where to put the bishop. ]
17...♖ae8 18.♕c2 ♗h4?! QUESTION: Why
not take on c3?
[ ANSWER: Good question! That was
almost certainly the best move. Bronstein
correctly indicates 18...♗xc3 19.♘g5
as the best reply, saying that "at the board,
it was difficult [for Black] to calculate all
the consequences of this move". I don't
think a calculator of Korchnoi's class
would have found it so difficult, actually.
After ♕f5 Bronstein gives
A) instead, White should prefer 20.♕b3+
♔h8 21.bxc3 ♘a5 ( 21...♕xf4?
22.♘e6 ) 22.♕b1 ♕xb1 23.♖axb1 ♖xf4
24.♖fe1 , with a small advantage, but it is
not a great deal;
B) 20.♕xc3 ♕xf4 21.♘xh7
, but this in fact loses after ♘d4 22.♕d3
♖f5 and the white knight is trapped.
QUESTION: So why didn't Korchnoi
play this?
ANSWER: I don't know! I suspect he
just misjudged the position reached in the
game. ]
19.♗g3
[ 19.♘xh4 ♖xf4 20.♘f3 ♖xf3 21.gxf3
♘e5 gives Black some counterplay,
although the computer remains
unconvinced that it amounts to much after
22.♕e4 ♘d3 23.♕d4 . ]
[ 19.♗e3 is another excellent move, but as
we have already seen, Bronstein preferred
to have his bishop more actively placed,
attacking c7. ]
19...♗xg3 20.hxg3 QUESTION: But now
 Bronstein − Move
11 by Move
White's extra pawn is doubled!
ANSWER: True, and that may have been
what Korchnoi was counting on. But it is
still an extra pawn, and White retains a clear
advantage. The other factor we have not yet
seen come into the game is the potential
weakness of Black's queenside pawns, which
can be attacked with a4.
♘e5 21.♘xe5 ♖xe5 22.♖fe1 ♖d5
[ Bronstein criticizes this and prefers
22...♖fe8 , but then his line 23.♖xe5 ♖xe5
24.♖d1 ♕e7 25.♕d2 , with the intention
o f m e e t i n g ♖e2 with the awkward check
26.♕d5+ , still leaves White clearly
better.
QUESTION: It still seems strange to me
that White should be so much better here,
when he has only an extra doubled pawn.
ANSWER: The point is that Black's king is
also weaker than White's here, and in
heavy piece endings, king safety is a
critical factor. The rest of the game is a
very instructive illustration of this. ]
23.♖ad1 c5 24.a4! The key idea of
attacking the queenside pawns comes into
effect.
QUESTION: But I thought you just said it
was more about king safety here?
ANSWER: I did and it is, but, as always, the
key strategic principle of a second front
applies. Bronstein wants to weaken the
enemy queenside pawns, and then combine
threats against them with an attack on the
slightly more exposed enemy king. The
combination of the two is much more likely
to overwhelm the defence than a kingside
attack by itself.
♖d8
[ 24...bxa4? 25.♖xd5 ♕xd5 26.♕xa4
just weakens the black queenside pawns
even more, whilst the counterattack ♕d2
is easily repulsed: 27.♕c4+ ♔h8 28.♖e2
Bronstein − Move by Move
♕c1+ 29.♔h2 ♕h6+ 30.♕h4 . ]
25.♖xd5 ♕xd5 26.axb5 axb5 27.♕e2!
Attacking b5, as well as threatening to
penetrate to e7.
b4 28.cxb4 cxb4 29.♕g4 b3 30.♔h2!
A v i t a l l y i m p o r t a n t p r e p a r a t o r y mo v e .
Before starting his attack, White tucks his
own king into safety, thus emphasizing the
key element of his advantage – Black has no
comparable way to secure his own king.
[ I n s t e a d , t h e i m m e d i a t e 30.♖e7?
leads to nothing after ♕d1+ 31.♕xd1
♖xd1+ 32.♔h2 ♖d2 with a drawn rook
ending. ]
30...♕f7 Black prevents the rook entering e7.
[ Another defence was 30...h6 31.♖e7 ♕g5
, but White retains a clear advantage after
32.♕e6+ ♔h7 33.♕e4+ ♔g8 and now the
q u e e n e n d i n g a f t e r 34.♖e8+ ♖xe8
35.♕xe8+ ♔h7 36.♕e4+ ♔h8 37.♕d4
sees White retain a substantial plus,
although much work would still remain
ahead. ]
31.♕g5 ♖d7 32.f3 QUESTION: What is
happening? White seems to have achieved
little with his attack.
ANSWER: He has forced the black pieces
into passive positions, as they must prevent
the white rook penetrating to e8 and e7.
There is no immediate way to make progress,
so White slightly improves his king position,
getting the f2-pawn off prise and ensuring
that a later ...Rd2 (for example, after a queen
exchange) will only attack one pawn, not two.
This might prove vital later on. Meanwhile,
he invites Black to find a move.
h6 33.♕e3 ♖d8
[ T h e t h r e a t w a s 33...-- 34.♕e8+
, forcing a winning rook ending, since after
♕xe8 35.♖xe8+ ♔f7 36.♖b8 and now
♖d2 d o e s n o t h i n g a f t e r 37.♖xb3
, as the pawn is not attacked on f2 – see
 Bronstein − Move
12 by Move
the previous note! ]
34.g4 QUESTION: What does that do?
ANS WER: W h i te again makes a sli g h t
improvement in his kingside structure. Now
there is no check on h5, and his king has a
further escape square on g3, and can enter
the game more quickly, should queens come
off. Meanwhile, g4-g5 may later be a threat.
♔h8 35.♕b6! The start of a beautiful
winning idea, based on deep calculation.
White attacks the pawn on b3, and prepares
simply to win it with Re3. However, at the
same time, he allows Black a dangerous-
looking counterattack, so he had to have
seen everything that follows.
♖d2 Both attacking b2 and threatening 36...
Qxf3.
36.♕b8+ ♔h7 37.♖e8! ♕xf3? Falling into
the trap.
[ Black could defend much more
t e n a c i o u s l y w i t h 37...♔g6 , but then
simply 38.♔h3 removes the threat against
the f3-pawn, and retains all the advantages
of the white position.
QUESTION: But isn't b2 hanging?
ANSWER: No. After ♖xb2? White wins
by 39.♕d6+ ♔h7 ( if 39...♕f6? 40.♖e6;
or 39...♔g5? 40.♖e5+ ) 40.♕d3+ g6
( 40...♕g6 41.♖h8+ ) 41.♖d8
and the weakness of the black king is again
decisive – an absolutely typical scenario
for heavy piece endings. ]
38.♖h8+ ♔g6 But now it is not clear what
White has in mind.
EXERCISE: Can you find Bronstein's
winning blow?
ANSWER: 39.♖xh6+!!
[ 39.♖xh6+ is a beautiful blow, which
brought immediate resignation. Black loses
his queen after gxh6 ( or 39...♔xh6
40.♕h8+ ♔g6 41.♕h5+ ♔f6 42.g5+ ,;
whilst it is his king that is lost after
Bronstein − Move by Move
39...♔f7 40.♕c7+ ♔g8 41.♕c8+ ♔f7
42.♕e6+ ♔f8 43.♖h8#; or 39...♔g5
40.♕e5+ ♔xg4 41.♖g6+ ♔h4 42.♕g5# )
40.♕g8+ ♔f6 41.♕f8+ .
A beautiful combinative finish by
Bronstein, but also a very instructive
example of the main factors in heav y
piece endings, namely king safety and the
two weaknesses principle. ]
1-0
12 C86
Bronstein,D
Winiwarter,F
Krems (Game 3) 1967
[Steve Giddins]
1.e4 e5 2.♘f3 ♘c6 3.♗b5 a6 4.♗a4 ♘f6
5.0-0 ♗e7 6.♕e2
[ The Worral Attack, a nice change from
the stereotyped 6.♖e1 . ]
[ 6.♕e2 was always a favourite of Alekhine
and Keres, and amongst contemporary
GMs, Sergey Tiviakov is its main
proponent.
QUESTION: How is it different?
ANSWER: White often brings the rook to
d1, creating potential threats along the d-
file, thanks to the opposition of the rook
and the black queen.
QUESTION: But it doesn't really confer
any advantage against best play, does it?
ANSWER: No, but it changes the contours
of the regular lines, and at least avoids
the Marshall Attack, which has been
analysed to death. ]
6...b5 7.♗b3 d6
[ Black can also develop along pseudo-
Marshall lines with 7...0-0 8.c3 d5
w h e n W h i t e d o e s b e s t t o p l a y 9.d3
with chances that are objectively equal,
 Bronstein − Move
13 by Move

but with plenty of scope to outplay the temporary state of affairs. Meanwhile, White
opponent, ( as 9.exd5?! e4 10.dxc6 ♗g4 sets up a solid central structure, and will now
is quite dangerous. )] turn to the exploitation of the offside bishop
8.c3 ♗g4?! As discussed in Game 1, this on h5. This will be driven back to g6, by an
move is somewhat dubious before White has eventual g2-g4, and only then will White
advanced his pawn to d4. expand in the centre. White's advantage is
[ The usual move is 8...0-0 with equal not a large one, but it is stable and long-
chances. ] lasting. There have been many games in
9.h3 Diagram such structures, and these have shown that
XHGFEDCBAY Black suffers from his queen's bishop being
stuck out of play on g6.
1-mKR+-vLNtR! 0-0 11.♘bd2 ♘a5 12.♗c2 c5 13.♖e1
2+PzPQzP-zPP" 14 All thematic stuff from both sides. With his
last move, White freed the f1-square for his
3P+N+-zPL+# knight, and now intends Nf1, g4 and Ng3.
♘d7 Diagram
4+l+P+-+-$
5-+-zp-+p+% XHGFEDCBAY
6+-sn-zpn+p& 1-mK-tR-vL-tR![
7pzppvl-zp-+' 2+PzPQsNLzPP" 15
8tr-+kwq-+r({ 3P+N+PzP-+#
xhgfedcbay 4+-+P+-+-$
5l+-zp-zppsn%
♗h5!? QUESTION: The punctuation
suggests you are not entirely convinced by 6+-+-zp-+p&
this. 7pzppvln+-+'
ANSWER: True. The move is consistent, but
now the bishop is liable to end up out of play 8+ktr-wq-+r(
for some time to come. On the other hand,
retreating back along the c8-h3 diagonal, for
xhgfedcbay
e x a m p l e t o d 7 , w o u l d o b v i o u s l y b e a 14.g4! QUESTION: What is the idea of that?
significant concession, as White would have [ ANSWER: I think Winiwarter intended to
gained the useful move h2-h3 for nothing. meet 14.♘f1 with ♗xf3 15.♕xf3 ♗g5
10.d3! QUESTION: An exclamation mark , seeking to relieve his somewhat
seems a bit much for such a modest move. cramped position by a couple of exchanges
Hasn't Black achieved something, by forcing and, in particular, to offload his rather
White to abandon his planned central bad dark-squared bishop, which is hemmed
advance d2-d4? in by the black central pawns. However,
ANSWER: He has delayed it for the time Bronstein does not allow this. ]
being, but as we will see, this is only a 14...♗g6 15.♘f1 f6?! QUESTION: That

Bronstein − Move by Move

 Bronstein − Move
14 by Move
doesn't do much for his king's bishop!
ANSWER: No. If Winiwarter had followed
Jonathan Rowson's advice and "talked to his
pieces", I suspect the king's bishop would
have had some choice words to say about the
move 15...f6.
QUESTION: So what do you recommend?
[ ANSWER: I think 15...h6! is better, still
hoping to trade off his bad bishop with a
subsequent ...Bg5. White would retain
some advantage, but less than in the
game.
QUESTION: But is White better, even in
the game? Fritz does not seem to think so.
ANSWER: Yes, I noted that also. But I
think this is one case where the
computer's judgement must be taken with a
pinch of salt. Computers still have a lot of
trouble assessing such blocked and slow
positions. I repeat that White's advantage
is not enormous, but he is definitely for
preference here. Black's queen's bishop is
out of play, his king's bishop is bad, and
once the white pawn reaches d5, as it will
inevitably do, Black will have the usual
Chigorin system problems with his queen's
knight. ]
16.♘e3 ♗f7 17.d4 Here we go. As predicted,
White's restrained set-up with the pawn on
d3 was only a temporary expedient, and now
that he has completed the necessary
preparations, he expands in the centre.
♖e8 QUESTION: Could he exchange in the
centre?
[ ANSWER: It is also possible, yes, but
after 17...cxd4 18.cxd4 ♖c8 ( capturing
twice in these structures, with 18...exd4
is rarely good, unless Black can follow up
with a quick ...d6-d5 and/or exert pressure
against the white centre, with moves such
as ...Bb7, ...Re8, etc; here, his pieces are
all hopelessly misplaced for such a strategy,
Bronstein − Move by Move
and White would be much better after
19.♘xd4 ) 19.a4 White retains the typical
Lopez pull. ]
18.d5 QUESTION: That's a major decision,
isn't it?
ANSWER: Yes, it is. White abandons any
attempt to break through in the centre or
exploit the weakened d5-square, and instead
places his hopes on a flank breakthrough. In
fact, as will become clear later on , his
strategy is very deep. Nowadays, thanks
particularly to Karpov's influence, d4-d5 is
pretty much the standard approach to these
Spanish structures. Fischer was always
rather more keen on the Rauzer plan with
dxc5 and trying to exploit the d5-square, but
Karpov perfected the slow manoeuvring
approach to the structure.
c4 19.b4 ♘b7 See the note to move 15.
Now the knight is very bad on b7, which is a
typical problem for Black in these positions.
QUESTION: So why doesn't Black take on
b3?
[ ANSWER: The cure would be no better
than the disease. After 19...cxb3 20.axb3
White has pressure down the a-file against
the pawn on a6, whilst the knight still
lacks any decent retreat square from its
current offside position on a5. White can
continue with Bd2, setting up tactical
threats against the knight, with c3-c4. ]
20.a4 ♕c7 Diagram
(Diagram 16)
 Bronstein − Move
15 by Move
XHGFEDCBAY
1-mK-tR-vL-tR![
2+-zPQ+L+-" 16
3P+NsN-zP-+#
4+P+P+pzPP$
5-+-zpP+p+%
6+-zp-zp-+p&
7pzplvlnwqn+'
8+k+r+-+r(
xhgfedcbay
21.a5!? QUESTION: Hmm, that is a
su rp r i se! I t h ought White would t ry t o
exploit the queenside?
ANSWER: Yes, it is a surprising decision,
and I am not 100% sure it is best. In his book
'200 Open Games', Bronstein presents this
game via an amusing, but, I suspect,
apocryphal story, in which he keeps leaving
the board and joining a friend amongst the
spectators, who is a Professor of Geology.
When Bronstein plays d5, the Professor
expresses his surprise at the decision to close
the centre. Then after a5, he is even more
surprised that Bronstein thinks he can break
through solely on the kingside. Finally, as
we will see, Bronstein closes the kingside as
well, leaving his friend totally bewildered by
the apparent decision to play for a draw.
Then Bronstein finally unveils the point of
his plan, which is to break things open with a
sacrifice. When events pan out exactly as
explained, the professor is suitably
impressed by the depth of the Grandmaster's
vision: "I think in chess, you can see deeper
than seven kilometres down!".
It is all very nice, but is it really a true
Bronstein − Move by Move
reflection of what happened? It is hard to say.
Certainly, I do not see any real need to close
the queenside. White could just double rooks
on the a-file, when the threat of axb5 would
pretty much force Black to cede the file. Of
course, White still has to find a way to break
through, but that would be the standard plan.
g6 22.h4 With the centre and queenside
closed, Bronstein sets about preparing a
break on the kingside.
♔g7 23.♔g2 h6 24.♖h1 ♖h8 25.h5
QUESTION: But surely this just closes the
kingside up?
ANSWER: You are starting to sound like a
Soviet geology professor! This is the point in
'200 Open Games', where Bronstein finally
reveals his idea, so we may as well do the
same ourselves. White's plan is to sacrifice a
piece on c4, and then set in motion a pawn
roller. Similar plans frequently occur in
King's Indian and Czech Benoni structures,
where White has pawns on c4-d5-e4 against
Black pawns on c5-d6-e5. Then, too, White
often breaks through by sacrificing a piece
on c5. A classic example is the famous
correspondence game Berliner-Nyman, from
the 1968 world correspondence
championship.
QUESTION: But surely, even if White
manages to sacrifice on c4, which Black can
try to prevent, it is far from clear that he
will be winning the resulting position?
ANSWER: Well, as to the first point, one of
Bronstein's arguments to the professor was
that Black cannot actually stop the sacrifice
on c4, because White can attack the square
with more pieces than Black can bring to its
defence. And, as to the second point, it is
true that White is not likely to be able to
calculate everything out to a forced win, but
his intuition tells him the position should be
winning. If we imagine the black c- and b-
 Bronstein − Move
16 by Move
pawns gone, and White's pawns moving up to
b4, c4 and then c5, one can see that Black's
position will be extremely hard to defend.
Th e acute lack o f s pace f or h is p ieces ,
especially the way the knight on b7 is under
its colleagues' feet, makes it very hard to
defend.
g5 26.♘f5+ ♔f8 27.♗e3 ♖h7 28.♘d2
Diagram
XHGFEDCBAY
1R+-+-+-tR!
2+KzPQsNL+-" 17
3-+-vL-zP-+#
4+P+P+pzP-$
5PzpNzpP+pzP%
6zp-zp-zp-+p&
7r+lvlnwqn+'
8+-mk-+-+r({
xhgfedcbay
White's next stage is to bring his pieces into
position for the sacrifice on c4. He proceeds
to do this systematically.
♗g8 29.f3 ♗d8 30.♗f2 ♗e7 QUESTION:
Black just seems to be shuffling his pieces
aimlessly.
ANSWER: Yes, because he has no
constructive moves at all. He can only await
the blow falling and then hope to be able to
deal with the tactics that result.
31.♖hc1 ♗d8 32.♕f1 In contrast with his
opponent's, White's play is extremely
systematic. His rook stands behind the c-
pawn, ready to support an eventual c4-c5.
Now he just needs to transfer his Spanish
bishop to e2 and bring back the knight from
f5 to e3, and all will be set.
Bronstein − Move by Move
♗f7 33.♗d1 ♗e8 34.♗e2 ♖c8 35.♘e3
The last brick in the wall. Note how White
left this move until last, maintaining the
knight on f5 as long as possible, where it
attacked the h6-pawn and so prevented
Black bringing his rook on h7 over to help
on the queenside.
♘b8 Diagram
XHGFEDCBAY
1-+Q+-tR-tR![
2+KvLLsN-+-" 18
3-+PsN-zP-+#
4+P+P+pzP-$
5Pzp-zpP+pzP%
6zp-zp-zp-+p&
7r+-+-wqn+'
8+-mklvlrsn-(
xhgfedcbay
36.♘dxc4! The breakthrough finally comes.
Black may as well capture, else the knight
just moves away again, and White plays c4
without even having invested any material.
bxc4 37.♘xc4 ♗b5 Losing the exchange,
but even the computer confirms that the
position is hopeless for Black.
[ A f t e r a m o v e s u c h a s 37...♕f7
, for example, meeting the threat of Nb6,
W h i t e c a n p l a y s i m p l y 38.♘e3
. Now a6 hangs, and cannot be defended.
Once that goes, the white queenside pawns
will be unstoppable. Note how the black
pieces are helplessly stumbling over each
other – the knight on b7 does not have a
single move, for example. ]
38.♘b6 ♗xe2 39.♕xe2 ♗e7 40.♘xc8
♕xc8 41.♗a7 The simplest. White does not
 Bronstein − Move
17 by Move

even bother with c4-c5, but just goes after

to catch his opponent in some trappy
the a6-pawn, when his own a-pawn will
variations (as happens in the present game),
decide. but objectively, it is insufficient to equalize.
♘d7 42.♕xa6 A very nice example of a
QUESTION: Bronstein seemed inordinately
slow manoeuvring struggle, and of the use of
fond of such dubious opening lines. How did
a piece sacrifice to break open these blocked
he get away with it?
structures. ANSWER: Well, of course, opening
1-0 knowledge was far, far more limited in those
days. There were no databases, and until the
late 1960s, not even any 'Informators'. The
13 C68 overwhelming majority of games played in
Nunn,J GM events never saw the light of day, except
Bronstein,D possibly in the official bulletins, which
Hastings (Game 4) 1975 would only be produced in a small number
[Steve Giddins] of copies and distributed to the players, plus
a handful of journalists. It was possible for a
1.e4 e5 2.♘f3 ♘c6 3.♗b5 a6 4.♗xc6 GM to play a certain opening line, and still
Diagram be catching opponents unawares a year or
XHGFEDCBAY two later, something which nowadays is
unthinkable, in the days of computer
1R+-mKQvLNtR! databases and live online game broadcasts.
2zPPzP-zPPzPP" 19 6.♘xe5 With Black planning a convenient
defence of his e5-pawn by 6...Ng6, this is
3-+N+-+-+# clearly the critical test of Black's 5th move.
Now, in order to regain his pawn, Black
4+-+P+-+-$ must expose his queen.
5-+-zp-+-+% ♕d4 Diagram
6+-+-+L+p& (Diagram 20)
7pzpp+pzpp+'
8trnvlkwql+r({
xhgfedcbay
The Exchange Spanish, a part-time favourite
of both Lasker and Fischer.
dxc6 5.0-0 ♘e7?! QUESTION: That looks
strange. What is going on?
ANSWER: This is one of the lesser-known
defences to the Exchange, but was something
of a hobby-horse of Bronstein's. It leads to
long forcing lines, where Black has chances

Bronstein − Move by Move

 Bronstein − Move
18 by Move
XHGFEDCBAY
1-mKR+QvLNtR![
2zPPzP-zPPzPP" 20
3-+-+-+-+#
4+-+Pwq-+-$
5-+-sN-+-+%
6+-+-+p+p&
7pzppsn-zpp+'
8tr-vlk+l+r(
xhgfedcbay
7.♕h5!
[ 7.♘f3 ♕xe4 reveals another point of 5...
Ne7 – the e-file is blocked, so Black is not
losing his queen after 8.♖e1 . But the text
move is stronger anyway, as White
activates his own queen and threatens the
f7-pawn. ]
7...g6 8.♕g5?! At the time this game was
played, almost nothing was known about 5...
Ne7, at least by the general public – there are
conflicting stories about how much of the
present game was Soviet opening analysis,
known to Bronstein. It was claimed in
several magazines that this was the case, but
John Nunn confirms to me that Bronstein did
not say anything to this effect after the game.
On the other hand, after thinking for 20
minutes (!) over 1...e5, he played the rest of
the game quickly, so he may have known it
all.
[ Largely on the basis of this game, theory
now recommends as stronger the line 8.♘f3
♕xe4 9.♕a5! attacking c7. This is
regarded as giving White the advantage,
although the theory runs quite deep. As it
happens, Bronstein did once face this
Bronstein − Move by Move
variation, in a game which I happened to
witness, at the 1991 Hastings Challengers.
His opponent was the young American GM,
Ilya Gurevich (now, alas, lost to the game
of chess). The game continued b6 10.♕c3
♖g8 11.♖e1 ♕d5 12.♘g5 ♗g7 13.♕g3
( 13.♕e3! may be even stronger) 13...♕d6
14.♕f3 ( 14.♕b3 is also good) 14...♗f5
15.d3 ( 15.♘c3! ) 15...h6 16.♘e4 ♕d7
17.a4 a5 18.♘a3 0-0-0 and Black had
solved most of his problems.
Bronstein went on to outplay his young
opponent in impressive style, but then
missed several wins and eventually lost.
Later that evening, I saw a thoroughly
disgruntled Bronstein having dinner with
friends in a local restaurant, and
complaining vociferously. His main casus
belli appeared to be that, after the game,
his opponent had started telling him where
he had gone wrong, instead of praising him
for having played so well for a man of his
age. ]
8...♗g7 9.♘f3 ♕xe4 10.♖e1 ♕b4
QUESTION: It looks as though Black's
position is hanging by a thread!
ANSWER: It is! He relies on only moves at
various points, but if he can survive and get
developed, then he will have his bishop-pair
and decent chances in the middlegame.
11.b3? QUESTION: Nunn's move looks
extremely dangerous. Why is it wrong?
ANSWER: Well, basically, it does not quite
work. At first glance, one would think Black
is just getting massacred, as the threat of 12
Ba3 seems lethal, but it transpires that he can
defend.
[ White should probably just settle for a
d e v e l o p i n g m o v e , s u c h a s 11.♘c3
, although he has no real advantage after
♗e6 .
QUESTION: Was Nunn making this up
 Bronstein − Move
19 by Move

over the board? 16.♕f1 0-0-0 is pretty horrible for White. ]

ANSWER: Yes. He was extremely unlucky, 15...♕xe2 16.♖xe2 0-0-0 QUESTION: But
in fact, as he comes very close to refuting why is White losing here? He seems to have
the whole black set-up, but in the end, it repulsed the threats against his king.
turns out that this is a false hope. What isANSWER: Yes, but the instability of his
more surprising is that this same mistake knight on e5 will cost him material. The
11 b3? has since been repeated by a number knight is pinned, and ...Nf4 is in the air. As
of very strong players, including Arkady we will see, White cannot hold things
Naiditsch. ] together, a fact that the computer confirms.
11...h6 12.♕e3 ♗e6 13.♗a3? Diagram 17.♗b2 QUESTION: Is there really nothing
[ ANSWER: Yes, but, surprising as it may
better?
seem, it is also t h e losing move. Th e ANSWER: Well, I am happy for you to
computer recommends settling for 13.♘c3 suggest something!
, a l t h o u g h a f t e r ♘f5 14.♕e2 0-0 [ QUESTION: OK, what about 17.d4 then?
Black is already somewhat better. It is ANSWER: In that case, I win immediately
ironic to see how it is the white queen with ♘f4 18.♖e4 ♖xd4! with an extra
which is being chased around by the pawn, whilst the knight on e5 is still in
enemy pieces, rather than Black's. ] trouble. ]
[ QUESTION: Hmm. Well, in that case, let's
XHGFEDCBAY try 17.c3 . At least that breaks the pin, as
well as preparing d4.
1-mK-tR-+NtR! A N S W E R : I n d e e d , b u t a f t e r ♖he8
2zPPzP-zPP+P" 21 you are in danger of being pinned down the
e-file instead, and if the knight moves away,
3-+NwQ-+PvL# say with 18.♘f3 , t h e n s i m p l y g5
gives Black a huge initiative. The white
4+-+-+-wq-$ queenside pieces on a1 and b1 cannot get
5-+-+-+-+% out, and meanwhile, Black has all his
pieces in play, and is threatening ...g4, ...
6zpp+l+p+p& Nf4, ...Bd5, etc. White is just not going to
7-vlpsn-zpp+' survive this. ]
17...♖he8 The same problem as in the
8tr-+k+-+r({ previous variation. Now there are potentially
xhgfedcbay two pins, one along the a1-h8 diagonal, and
the other down the e-file.
QUESTION: Why is this wrong? It is the 18.d3
consistent follow-up to his 11th move, isn't [ This loses at once, but after 18.♘c4 ♗xb2
it? 13...♘d5! Highly embarrassing. 19.♘xb2 ♗f5 20.♖xe8 ♖xe8 21.♘a3 b5
14.♕e2 White is once again tied hand and foot. ]
[ 14.♗xb4? ♘xe3 loses the exchange. ] 18...♘f4 19.♖e3
14...♕g4 15.♘e5 [ 19.♖e4 runs into ♘xg2! 20.♔xg2 ♗d5
[ Practically forced, since 15.♘c3 ♘f4 , with the additional point that after 21.♘c4

Bronstein − Move by Move

 Bronstein − Move
20 by Move

there is ♗xb2 22.♘xb2 f5 23.c4 (forced) XHGFEDCBAY

♗xe4+ 24.dxe4 ♖xe4 and Black has rook
and two pawns against the two knights, 1-mK-+-+NtR!
with the entire white army bunched up 2zP-zP-+PvLP" 22
hopelessly on the queenside. For example,
25.♘c3 ♖g4+ 26.♔f1 ♖d2 forces the 3-zP-tRP+P+#
knight to the edge of the board by 27.♘ba4
and Black's initiative remains too strong.
4+-sn-+-+-$
QUESTION: Well, I can see things are 5-+-sNl+-+%
pretty bad for White, but is it really
hopeless? 6zpp+-+p+p&
ANSWER: You want proof? OK, here is a 7-vlp+-zpp+'
sample line, offered by the computer: ♖h4
28.♔g1 ♖f4 29.♘d1 b5 30.cxb5 cxb5 8+-+rtrk+-({
31.♘c5 ♖fd4 32.♘e3 f4 33.♘g4 ♖c2
34.♘e6 ♖e4 35.♘f8 h5 36.♘f6 ♖ee2
xhgfedcbay
and the white forces have been scattered to 20...♘g2!
the four winds, whilst Black's rooks are [ After 20...♘g2 21.♖e2 ♗f3! material is
ready to start feasting on the enemy lost: 22.♘xf3 ♗xb2 23.♖xe8 ♖xe8
pawns. ] 24.♔xg2 ♗xa1 and Black's extra exchange
19...♗d5 20.g3 Diagram wins easily.
[ Or 20.♘c4 ♗xb2 21.♘xb2 ♘xg2 A remarkable game in its way. White's
22.♖xe8 ♖xe8 which is similar to the last early attempt to refute the black set-up
two notes. ] went awry, after which his pieces were
driven back in disarray, and he never
succeeded in untangling. ]
(Diagram 22) 0-1

14 C58
Bronstein,D
Rojahn,E
Moscow Olympiad (Game 5) 1956
[Steve Giddins]

1.e4 e5 2.♘f3 ♘c6 3.♗c4 ♘f6

[ As this game shows, Bronstein was a fount
of remarkable opening ideas, and this might
be an appropriate time to mention another,
which I came across some years ago in an
old issue of the Russian magazine,
'Shakhmaty v SSSR'. In the main line of

Bronstein − Move by Move

 Bronstein − Move
21 by Move
the Giuoco Piano, a well-known tabiya is
reached after 3...♗c5 4.c3 ♘f6 5.d4 exd4
6.cxd4 ♗b4+ 7.♗d2 ♗xd2+ 8.♘bxd2 d5
9.exd5 ♘xd5 10.♕b3 .
A) Now the best move is generally agreed
to be 10...♘a5 , but this has the
significant drawback that White can force
a d r a w a f t e r 11.♕a4+ ♘c6 12.♕b3
, etc.;
B) Black players wishing to fight for a
win here are therefore forced to resort to
10...♘ce7 , which avoids the immediate
draw, although White probably retains a
pull, objectively.;
C) I for one certainly never seriously
entertained the thought that Black might
have any other playable move in this
position, but I had failed to reckon on the
tireless creativity of David Bronstein. In
a clock simultaneous in the 1960s, against
the well-known endgame study composer
Ernest Pogosyants, Bronstein played the
remarkable move 10...0-0!? .
It looks like madness, but, on closer
i n s p e c t i o n , a f t e r 11.♗xd5 ♘a5
B l a c k r e g a i n s t h e p i e c e . 12.♗xf7+
( in the game, Pogosyants refrained from
taking the pawn, and was holding the
balance until very near the end; the rest
of the game is not of great importance in
itself, but once again, I offer the
complete score, for the benefit of
collectors of Bronstein game scores:
12.♕c3 ♕xd5 13.0-0 ♘c6 14.♖fe1
♗e6 15.♘b3 ♕h5 16.♘e5 ♗d5
17.♘xc6 bxc6 18.♕g3 a5 19.♘c5 ♖fe8
20.♖e5 ♕g6 21.♕xg6 hxg6 22.♖ae1
♖xe5 23.♖xe5 ♖b8 24.h4 ♖xb2
25.♖e8+ ♔h7 26.♘d7 g5 27.hxg5
♔g6 28.♘e5+ ♔xg5 29.f4+ ♔f6
30.♘g4+ ♔f5 31.♘e3+ ♔xf4
32.♘xd5+ cxd5 33.♖e7 f5 34.♖xg7
Bronstein − Move by Move
♖xa2 35.♖e7 ♖a4 36.♖e2 ♖xd4
37.♔f2 a4 38.♖a2 ♔e4 39.♔e2 f4
0-1, E.Pogosyants-D.Bronstein) 12...♖xf7
is then critical, when Bronstein claims
that Black has sufficient initiative to
compensate for the pawn. That is
perhaps objectively questionable,
although Black's position does not looks
so much worse than in the main line of
the 4 Ng5 Two Knights.;
D) 10...-- ]
4.♘g5 QUESTION: Is this good? Tarrasch
dismissed it as bad, didn't he?
ANSWER: Yes, Dr. Tarrasch famously
derided it as "a duffer's move", whereas
Steinitz stubbornly stuck to it, despite losing
a number of bad games against Chigorin.
The dispute rumbled on for decades, but the
advent of computers has certainly swung
grandmaster opinion towards the white side,
and nowadays, one finds few GMs willing to
play the black side of the Two Knights.
d5 5.exd5 ♘a5 And this is the main reason
why. Not only does Black lose a pawn, but he
is also forced to put his knight on the edge of
the board. Although White is behind in
d e v e l o p me n t , h i s p o s i t i o n c o n t a i n s n o
weaknesses at all, and nowadays, strong
players tend not to trust such sacrifices.
[ Unfortunately for Black, the two
alternative continuations 5...b5 ]
[ and 5...♘d4 , which see the black knight
occupy a central post, have both been
analysed out to a white advantage. ]
6.d3?! A very unusual choice for the 20th
century.
[ The line that has put people off the Two
K n i g h t s i s 6.♗b5+ c6 7.dxc6 bxc6
a n d n o w b o t h 8.♕f3 ( and Steinitz's old
move 8.♗d3 have achieved excellent
results for White in the computer era. )]
[ QUESTION: But is there anything wrong
 Bronstein − Move
22 by Move
w i t h B r o n s t e i n ' s 6.d3 ? It looks quite
logical to defend the bishop and open the
other white bishop's line.
ANSWER: It is, but the problem is the
sequence which follows. ]
6...h6 7.♘f3 e4 This is the rub.
[ For many decades, it has been assumed
t h a t a f t e r 7...e4 White has nothing better
than 8.♕e2 , but after ♘xc4 9.dxc4 ♗c5
Black has excellent play for the pawn.
However, for this game, Bronstein had
prepared a very special idea. ]
8.dxe4!? Astounding.
QUESTION: Err, doesn't this...?
ANSWER: Yes, it does. White is giving up a
whole piece, for which he will have only a
couple of pawns, plus a central pawn duo. At
first sight, it looks like madness, and indeed,
the line has never been trusted and has not
been repeated at GM level, but it works on
this occasion.
QUESTION: What was the reaction of
Bronstein's teammates?
ANSWER: Ah, well thereby hangs an
interesting tale. This game was played in an
Olympiad. The Soviets had a very good rule
in operation within their team, to the effect
that nobody was allowed to play in a manner
that mig ht caus e their teammates to b e
alarmed about their position. This was for
the very good pragmatic reason that, in a
four-board match, if the other players start to
believe that teammate X is losing, they are
liable to start taking liberties in their own
games, to try to generate winning chances to
make up for him. This very often makes the
match situation worse, and so the Soviets
al wa y s t r i ed t o av o i d s u ch d an g e r s . I n
accordance with this 'rule', Bronstein records
that he had shown his idea to his teammates
the night before, and had received their
blessing to his trying it out.
Bronstein − Move by Move
QUESTION: But his opponent obviously did
not know about this?
ANSWER: No, According to Bronstein, after
the game, his opponent complimented him on
his coolness and cold-bloodedness, in calmly
playing on and finding the best chance, after
"blundering" away a piece!
♘xc4 9.♕d4 ♘b6?! This is already a
critical moment.
[ Rojahn's choice is the most obvious, but
in actual fact, 9...♘d6 was more critical.
QUESTION: But that just sets the knights
up to be forked, doesn't it?
ANSWER: It looks that way, but tactical
resources change the assessment, and after
10.♘c3 Black has the nasty trick ♘fxe4
11.♘xe4 ♕e7 , when he regains the piece
and eliminates one of White's central pawns.
After the further moves 12.0-0 ( 12.♘e5
♘xe4 13.♕xe4 f6 14.f4 fxe5 15.fxe5
♕b4+ is also better for Black) 12...♘xe4
( and not 12...♕xe4?? 13.♖e1 ) 13.♖e1 f5
it looks a bit shaky, but Black should
emerge with most of his booty intact: for
example, 14.♘d2 ♕c5 15.♕e5+ ♔f7
16.♘xe4 fxe4 17.♕xe4 ♗d7
and, objectively, White does not have
enough.
QUESTION: So is 9...Nd6 just a
refutation? What did Bronstein intend?
ANSWER: I am not sure! It is curious
that neither in '200 Open Games', nor
'The Sorcerer's Apprentice', does
Bronstein mention the move 9...Nd6, or say
what he had in mind. ]
10.c4 c5? Black was clearly totally thrown
by events on the board, and it is perhaps not
surprising that he found it hard to orient
himself. The mass of white pawns from c4 to
e4 looks dangerous, and Rojahn hastens to
blockade the c-pawn, but the result is that
the white d-pawn is now passed, whilst
 Bronstein − Move
23 by Move
Black's king's bishop has limited scope.
[ He would have done better to play
10...♘bd7 followed by 11...Bc5.
QUESTION: Wasn't he frightened of 11.e5
? I know I would be.
ANSWER: It does look a bit scary, I agree,
but the computer shows that ♗c5 12.♕h4
( 12.♕d3?! ♘g4 ) 12...♕e7 is better for
Black, as the white centre is overextended.
For example, 13.♗f4 ♘xe5 14.♗xe5 g5
15.♕g3 ♘h5 wins the queen. However,
one should not be too critical of Rojahn, as
it is all too easy to sit in the comfort of
one's study, with Fritz whirring away on
one's computer, and think how easily
Black could have refuted the sacrifice.
Over the board, with the clock ticking and
the other players crowding round the board,
looking at Bronstein's latest amazing idea,
it is much, much harder to keep one's cool
and analyse the tactics accurately. One
must also bear in mind the God-like
reputation that the leading Soviet players
had in that era – I should imagine this
was such that, when Bronstein played 8
dxe4, his opponent will have 'believed' him,
and already mentally accepted that he had
run into a powerful new idea. In such a case,
it is very hard to maintain one's
objectivity and play normally. ]
11.♕d3 ♗g4 12.♘bd2 ♗e7 13.0-0
Diagram
(Diagram 23)
Bronstein − Move by Move
XHGFEDCBAY
1-mKR+-vL-tR!
2zPPzP-sN-zPP" 23
3-+N+Q+-+#
4+l+P+P+-$
5-+-+Pzp-+%
6zp-sn-+-sn-&
7-zppvl-+pzp'
8tr-+kwq-+r({
xhgfedcbay
QUESTION: So now what is the assessment?
White is still a piece down and he does not
seem to have any threats.
ANSWER: True, but Bronstein was ready for
that. He is placing his faith in the long-term
advantage of the mobile pawn centre. This is
very much a case of what Rudolf Spielmann
termed a 'real sacrifice', where there is no
immediately calculable way to regain the
material or give mate. Rather, it is a long-
term sacrifice, where White hopes his two
pawns and massive pawn centre will offer
sufficient compensation for the piece.
0-0!?
[ Bronstein later criticized this, and said
Black should have taken his final chance to
exchange off this bishop for the knight with
13...♗xf3 . Now it becomes a target for the
onrushing white pawns. ]
14.♘e5 ♗h5 15.b3 ♘bd7 16.♗b2 ♘xe5
17.♗xe5 ♘d7 18.♗c3 ♗f6 QUESTION:
Black's play looks very logical!
ANSWER: It does and it is. He has
exchanged off his badly-placed knight on b6
and now seeks the exchange of dark-squared
bishops, whilst at the same time maintaining
 Bronstein − Move
24 by Move
control of e5. The computer still assesses the
position in Black's favour.
19.♖ae1 ♗xc3 20.♕xc3 ♕f6
[ 20...f6 is the computer's choice here, but
after 21.f4 play continues along similar
lines. Fritz still claims a small edge for
Black, but its assessment is dropping
with every move, which usually means that
the players knew what they were doing. ]
21.e5
[ QUESTION: Wait a minute! Isn't 21.♕h3
attacking two pieces?
ANSWER: Yes, but Rojahn had seen this
and intended the reply ♕b2 . If then
22.♘b1 ♗e2 23.♕xd7 ♗xf1 24.♖xf1
♕xa2 and Black has returned the exchange,
to neutralize the white initiative.
QUESTION: So you mean Black was
trying to equalize?
ANSWER: So it would seem, and
Bronstein's response shows that he was
not interested in such a continuation.
Although Fritz still thinks Black is better,
it is clear that both players were of a
different opinion during the game. ]
21...♕f5 22.f4 Now the pawn centre really
starts to assume menacing proportions, and
even the silicon beast agrees that White is
now very much on top. Bronstein's faith in
his position has been justified, and the
remainder of the game is pure pleasure for
White.
♗g6 23.♘e4 ♖ab8 QUESTION: What is
the point of that?
ANSWER: Black just defends the b7-pawn
against the threat of 24 Nd6. But such a
passive move can hardly be good.
[ 23...b6 looks like a better way to do the
same thing, although it does not change
the course of the game. ]
24.♕f3 Now g4 is also on the agenda.
♗h7
Bronstein − Move by Move
[ 24...♘xe5 was a try to disrupt the course
of White's build-up. After 25.fxe5 ♕xf3
( or 25...♕xe5 26.♘xc5 when "White has
an extra passed pawn" – Vainstein) 26.♖xf3
♗xe4 27.♖xe4 "the double rook ending
with an extra pawn is won for White"
(Vainstein). ]
[ 24...♗h5 is a better option, also mentioned
b y V a i n s t e i n , w h o t h e n g i v e s 25.♘g3
( Fritz's simple line 25.♕e3 ♗g6 26.♘d6
♕d3 27.f5 ♕xe3+ 28.♖xe3 looks even
more crushing) 25...♗xf3 26.♘xf5 ♗g4
27.♘e7+ ♔h8 28.f5 and claims a winning
advantage for White. ]
[ Finally, there is 24...♕h5 , which the
computer suggests as the best chance. After
25.♕xh5 ( 25.g4 ♗xe4 26.♖xe4 ♕g6
27.f5 ♕g5 is less clear, according to the
beast ) 25...♗xh5 26.f5 White retains what
should be a decisive advantage, although
some work remains to be done. ]
25.g4 ♕g6 26.f5 Now it turns into a
massacre, as the pawn armada sweeps away
all resistance.
♕b6 27.♕g3 f6 28.e6 If anything justifies
Bronstein's original idea of 8 dxe4, it is this
position.
♘e5 29.h4! It seems only fair to include
White's eighth and final pawn in the fun.
♔h8 30.g5 ♖bc8 31.♔h1 QUESTION:
Huh?
ANSWER: Yes, I know it looks a bit strange.
But besides setting up gxf6 and then Rg1,
there is probably also an element of sadism
about it – Black is invited to consider that,
if White has time for such moves, perhaps it
is worth considering resignation.
♕d8 Not on your life!
32.g6 ♗xg6
[ Unable to face the horrors of 32...♗g8
33.d6 Black makes a desperate sacrifice,
but it does not long delay the inevitable. ]
 Bronstein − Move
25 by Move

33.fxg6 b5 34.d6 ♕b6 35.d7 ♘xd7 15 C32

[ After 35...♖b8 White has many ways to Bronstein,D
win, but I suspect Bronstein might have Tal,M
chosen 36.♕xe5 ( 36.e7 is simplest of all) USSR Team Championship, Riga (Game 6)
36...fxe5 37.♖xf8+ ♖xf8 38.e7 . ] [Steve Giddins]
36.exd7 ♖cd8 Diagram
XHGFEDCBAY 1.e4 e5 2.f4 It is safe to say that the King's
Gambit did not appear too often on the top
1K+RtR-+-+![ boards of the USSR Team Championship.
2+-+-+-+P" 24 But when it did, Bronstein or Spassky were
usually involved, and the two both took
3-wQ-+-+P+# some notable scalps with it, at this highest of
levels.
4zP-+N+P+-$ d5 True to his own style, Tal offers a
5-+-+-zpp+% counter-gambit, but he runs into the modern,
rather scientific treatment of this most
6zpPzp-+-wq-& romantic of lines.
7-zp-+P+-zp' 3.exd5 e4 4.d3 ♘f6 Diagram
8mk-tr-tr-+-( XHGFEDCBAY
xhgfedcbay 1RsNLmKQvLNtR![
37.♘xf6! A final sacrifice to finish off a
2zPP+-+PzPP" 25
masterpiece. 3-+-+P+-+#
♕c6+ 38.♕g2 One of Bronstein's most
famous games. The remarkably imaginative
4+-zPp+-+-$
piece sacrifice 8 dxe4 does not appear to be 5-+-+P+-+%
really correct (especially after 9...Nd6!), but
one must remember that Rojahn was no fool 6+-sn-+-+-&
– he was good enough to get in the powerful 7pzpp+-zppzp'
Hungarian Olympiad team, for a start. The
game illustrates how effective such long- 8tr-vlkwqlsnr(
term sacrifices can be in an over-the-board
situation, with the defender taken by surprise
xhgfedcbay
and having limited time on the clock. 5.dxe4! This move and the next are the
1-0 modern treatment, which has put many black
players off the Falkbeer Counter-Gambit.
♘xe4 6.♘f3! This is now firmly established
as the best line for White, with the present
game being one of those which demonstrated
this. It leads practically by force to an
endgame where White has the better chances.

Bronstein − Move by Move

 Bronstein − Move
26 by Move
QUESTION: An endgame! But surely that is
not why White plays the King's Gambit, is it?
ANSWER: Well, it certainly was not
originally why people played the King's
Gambit, but starting even from Rubinstein's
analyses in the 1920s, masters begun to
realize that the romantic sacrifices of the
19th century were all very nice to see, but
were not objectively that strong, in many
cases. On the other hand, the exchange of
White's f-pawn for the enemy e-pawn, and
the subsequent exchange of the e4-pawn for
the black d-pawn, leads to a structure where
White has a potentially mobile queenside
pawn majority, which can often make the
endgame better for him. Many of the games
Spassky and Bronstein won in the King's
Gambit were based on reaching a superior
endgame.
[ The older move 6.♗e3 is seen in Game
8. ]
6...♗c5 Diagram
[ John Shaw's book on the King's Gambit
suggests that 6...♗e7 ]
[ or 6...c6 are better practical options,
although neither really offers quite
sufficient compensation for the pawn. ]
(Diagram 26)
Bronstein − Move by Move
XHGFEDCBAY
1R+LmKQvLNtR![
2zPP+-+PzPP" 26
3-+N+-+-+#
4+-zPn+-+-$
5-+-+Pvl-+%
6+-+-+-+-&
7pzpp+-zppzp'
8tr-+kwqlsnr(
xhgfedcbay
7.♕e2! The key idea in this variation. White
will force an exchange of queens on the e-file,
thanks to the instability of the enemy
knight's post on e4. He also sets up the move
Be3, putting paid to any threats against the
f2-square.
♗f5 Diagram
[ Q U E S T I O N : B u t s u r e l y 7...♗f2+
is good here, isn't it? It looks as though
White should just be losing quickly.
ANSWER: It seems so at first sight, but in
reality, this line is not dangerous. After
8.♔d1 ♕xd5+ 9.♘fd2! it is Black who is
in trouble, as his pieces are hanging. John
Shaw, whose magnum opus on the King's
Gambit is an indispensable reference work
to anyone interesting in the opening,
quotes the further moves f5 10.♘c3 ♕d4
11.♘cxe4 fxe4 12.c3! and the e4-pawn
drops, unless Black plays ♕e3 , whereupon
it is his own king which is in danger after
13.♕h5+ ♔f8 14.♗c4 . ]
[ QUESTION: OK, I see that. So, if Black's
problems stem from his knight on e4 being
unstable, what about 7...f5 defending it
securely?
 Bronstein − Move
27 by Move

ANSWER: Then White can continue 8.♗e3 move ago, since now d5 is defended as
disposing of the threat against f2. Once well. ]
again, we quote John Shaw's book: ♕xd5 9.♗e3 ♘xc3 QUESTION: Is this forced?
9.♗xc5 ♕xc5 10.♘c3 0-0 11.♘xe4 fxe4 [ ANSWER: No. Black has a choice of two
and now White again implements his similar endgames here. The other option is
strategy of heading for the better endgame: 9...♗xe3 10.♕xe3 ♘xc3 11.♕xe7+ ♔xe7
12.♕c4+! ♕xc4 13.♗xc4+ ♔h8 14.♘e5 12.bxc3 .
and he is just better, with the more active QUESTION: And how is this position?
pieces, better development and fewer ANSWER: Once again, White is better. For
pawn islands. ] the moment, he is a pawn up. Black can
regain the pawn on c2, but after ♗xc2
XHGFEDCBAY ( the main line is 12...♗e4 13.♘g5 ♗xd5
, but here, too after 14.0-0-0 , White's
1R+LmK-vLNtR![ superior development and initiative gives
2zPP+Q+PzPP" 27 Black problems; again, you should consult
a book on the King's Gambit for more
3-+N+-+-+# details, but for now, I ask you to trust me)
13.♔d2 ♗g6 14.♖e1+ ♔d6 15.♘d4
4+-zPn+-+-$ White has a strong initiative. ]
5-+l+Pvl-+% Tal chooses another sequence, also leading to
an ending where White will have the
6+-+-+-+-& superior development and some initiative.
7pzpp+-zppzp' 10.♗xc5 ♘xe2
[ 10...♕xe2+ 11.♗xe2 ♘xd5 also regains
8tr-+kwq-snr( the pawn, but then the further pawn
xhgfedcbay sacrifice 12.0-0-0! ♘xf4 13.♖he1
once again repeats the familiar scenario –
8.♘c3! White is fully mobilized, whilst Black has
[ QUESTION: Wait a minute. Isn't 8.g4 most of his pieces at home, and his king
winning a piece here? is exposed in the centre.
ANSWER: Yes, but it has been known QUESTION: It seems strange that king
since the 1920s that Black is much better safety should be such an issue, even with
after 0-0 9.gxf5 ♖e8 , as was the queens off the board.
demonstrated by a famous game ANSWER: Well, although I have been
Spielmann-Tarrasch, Mahrisch-Ostrau referring to these various positions as
1923. If you do not believe me, John "endgames", they are more in the nature
Shaw's book is again the place to look for of queenless middlegames. With so many
detailed analysis, as space considerations open lines and a large development
preclude us from going into the variations advantage for White, the traditional
more fully here. ] middlegame factor of king safety is still
8...♕e7 important, even in the absence of the
[ 8...♗f2+? 9.♔d1 is even worse than a ladies. ]

Bronstein − Move by Move

 Bronstein − Move
28 by Move
11.♗xe7 ♘xf4 12.♗a3 QUESTION: We
have another, very similar position to the last
note!
ANSWER: Yes. As we have seen, there are a
number of similar queenless positions which
can arise from this line, and in all cases, the
key factors are the same. Black can regain
his gambit pawn, and even nick a further
pawn, if he wishes, but White has a
significant lead in development, plus two
strong bishops, which are generally enough
to promise him the better chances. In fact, as
Bronstein acknowledges in '200 Open Games',
this very position was all old analysis by
Tartakower, in his 1920s book,
'Hypermodern Chess'.
♘d7
[ Tal decides against larceny, either by
12...♘xd5 13.0-0-0 , ]
[ or 12...♗xc2 13.♖c1 . ]
13.0-0-0 ♗e4
[ QUESTION: What is wrong with
13...0-0-0 here, getting the king out of
Dodge while he has the chance?
ANSWER: Amazingly, it loses a whole
piece!
EXERCISE: How?
ANSWER: 14.♖d4 and wherever the
knight moves, Black loses a piece: ♘g6
( or 14...♘h5 15.g4 ) 15.g4 and the bishop
on f5 is trapped in broad daylight! An old
friend of mine, a strong county player,
once fell into a similar trap against the
veteran English master Stuart Milner-Barry,
in a tournament in the late 1960s. As if
losing a piece in such a fashion were not
embarrassing enough, he then had to see
the game published in the BCM, with notes
by Harry Golombek, who described my
friend's play as "mildewed". ]
14.♘g5?!
[ The start of an imaginative sacrificial line,
Bronstein − Move by Move
but the computer suggests the simple 14.d6
i s v i r t u a l l y w i n n i n g f o r W h i t e : ♗xf3
15.gxf3 c6 ( or 15...cxd6 16.♗xd6 ♘e6
17.f4 and the two bishops slice Black's
position apart) 16.♖e1+ ♘e6 17.♗h3
, with a large advantage. ]
14...♗xd5 15.g3 This is the point of
Bronstein's idea. He was evidently
convinced that it was best, since he offers no
comment on it in '200 Open Games', but the
computer is sceptical.
♗xh1 16.gxf4 QUESTION: What has White
obtained for the exchange?
ANSWER: He has a fierce initiative, with
Bc4 coming next. But it looks as though
Black could have returned the exchange in
order to equalize.
c5?!
[ 16...0-0-0 is Fritz's choice, offering back
the exchange, to get the king out of trouble.
After 17.♘xf7 ♖he8 18.♘xd8 ♔xd8
it appears that White does not have very
much. ]
17.♗c4 ♗c6 18.♘xf7 Now the game reverts
to the lines Bronstein had envisaged. Despite
the simplified position, White has a very
powerful attack, which more than
compensates for the exchange.
b5 QUESTION: Is there anything else?
[ ANSWER: Not really. The only other
option worth considering is to chase the
bishop by means of 18...♘b6 , but then
19.♖e1+ ♔d7 20.♗e6+ ♔c7 21.♗xc5
and the threat of 22 Bd6+ forces Black to
surrender the exchange anyway. After
♖hd8 22.♘xd8 ♖xd8 23.f5 White has an
extra pawn, the two bishops and still a
strong initiative, and should win without
any great trouble. ]
19.♘d6+?! Another inaccuracy, which
imperils White's advantage.
[ The computer points out that simply
 Bronstein − Move
29 by Move
19.♗b3! retains all White's threats, since c4
20.♗xc4 bxc4 21.♘xh8 is much better for
him. ]
19...♔e7 20.♘xb5 ♖hf8? Now White is
again winning.
[ Black should have preferred 20...♗xb5
. Bronstein considered this, and gave the
variation 21.♗xb5 ♖hd8 22.f5 ♖ac8
23.b3 , concluding that "the pressure
ex e r t ed b y Wh it e's b is h o p s wo u ld b e
extremely unpleasant". This is perhaps
typical of pre-computer assessments. On
the face of it, his conclusion seems
perfectly valid, but Fritz just dismisses the
position as equal after, for example, ♘f6 . ]
21.♘d4! ♗g2 22.♘e6 ♖f5 23.♖g1 ♗e4
24.♘c7
[ Good enough, but Bronstein himself
pointed out that 24.♖e1! wins at once. He
admitted that he simply overlooked that,
after ♘f6 White has 25.♘xg7 , winning:
for example, ♖xf4 26.♘e6 ♖h4 27.♘xc5
♔d8 28.♘xe4 ♘xe4 29.♗d5 , etc. ]
24...♖d8 25.♖xg7+ ♔f6 26.♖f7+
Bronstein admitted that he actually thought
this was mate, and had missed the fact that
the black king can flee to g6!
[ H e i s s t i l l w i n n i n g , b u t 26.♖xh7
was simpler. ]
26...♔g6 27.♖e7 QUESTION: Is it clear that
White is winning? He has two pawns for the
exchange, but f4 and h2 both look weak.
ANSWER: The decisive factor is still the
strength of White's initiative. Now, for
example, the bishop on e4 is hanging and 28
Ne6 is also a threat, hence Black's next.
♘f6 28.♘e6 ♖c8 29.b3 A nice move,
defending the bishop on c4 and making room
for the dark-squared bishop to switch to the
long diagonal.
♖h5 30.♘g5 ♗d5 31.♗d3+ ♔h6 32.♗b2
Black is under constant harassment and still
Bronstein − Move by Move
does not have the time to take on h2 or f4.
The knight cannot move because of mate on
h7, so Black is driven to desperate measures.
c4 33.♗f5 c3
[ Black is losing a piece, in view of 33...♖e8
34.♗xf6 . ]
34.♗xc8 cxb2+ 35.♔xb2 ♖xh2 36.♖xa7
White has regained his exchange and has no
fewer than three extra pawns.
♖f2 37.♖a4 ♔g6 38.♖d4 h5 39.a4 h4
40.a5 ♗g2 41.a6 ♘h5 42.♗b7 ♘xf4
43.♖xf4
[ After 43.♖xf4 ♖xf4 44.♗xg2 ♔xg5
45.a7 the a-pawn costs Black a rook. It is a
pity that two errors could have cost
Bronstein his advantage, but overall, this
was an important game, both from a
theoretical viewpo int, and als o as an
illustration of the power of the initiative
in queenless middlegames. ]
1-0
16 C39
Bronstein,D
Ostrup
Gausdal (Game 7) 1994
[Steve Giddins]
1.e4 e5 2.f4 exf4 3.♘f3 g5
[ Bronstein's lifelong love affair with the
King's Gambit extended back to the
earliest days of his chess career. In the
'New in Chess' interview referred to in the
Appreciation, for example, he insisted at
one point on showing his interviewer a
nice little King's Gambit miniature that he
had won against an unnamed opponent, in
a simultaneous display in Kiev 1941:
3...♘f6 4.e5 ♘h5 5.♘c3 d6 6.♗c4 dxe5
7.♘xe5 ♕h4+ 8.♔f1 .
EXERCISE: Can you calculate a beautiful
 Bronstein − Move
30 by Move

finish? ( 9.♗g5 ♗e7 10.♕e2 was Mark Hebden's

ANSWER: ♗e6! (luring the white bishop
off the f1-a6 diagonal) 9.♗xe6 ♘g3+
10.♔g1 ♗c5+ 11.d4 ♗xd4+! 12.♕xd4
♘e2+! 13.♘xe2 ♕e1# . Bronstein does
not make clear whether he was the
simultaneous giver or not – presumably he
was, although it is somewhat unusual for
the exhibitioner to play Black. ]
4.h4 g4 5.♘e5 The Kieseritzky Gambit,
which is the nearest thing to a respectable
line after 3...g5. The main line frequently
leads to an endgame, where Black has an
extra pawn, but White has some positional
compensation. However, a few tests at top
GM level some years back, when the
Belorussian GM Alexey Fedorov ventured
the King's Gambit against such players as
Anand, Shirov and Ivanchuk, led to the
uncovering of some other defensive
possibilities for Black, which have rendered
the main line quite unattractive for the first
player.
♘f6
[ 5...d6 6.♘xg4 ♘f6 was the choice of 7...♘g3 Accepting the bait.
both Anand and Shirov and is an
effective equalizer. ]
6.♘xg4!? This is a pretty unusual move, and
one which has a poor reputation. Indeed, as
far as I can see, 6 Nxg4 is not even
mentioned in John Shaw's magnificent 8.♘d5
670-odd page magnum opus on the King's probably without justification.
Gambit. As we will see, Bronstein has a
creative idea in mind.
[ The two main moves are 6.d4 ]
[ and 6.♗c4 , but both have been very cold shower over this, with the simple
exhaustively analysed, which makes them
less attractive in these days of computer
preparation. ]
6...♘xe4 7.♘c3!?
[ The old line, going back to the Victorian ♔e6 16.♕e5# ) 15.♕h6+ ♔e7 16.♕e3+
era, was 7.d3 ♘g3 8.♗xf4 ♘xh1 9.♕e2+
creative attempt to improve the whole line
for White, but in Hebden-Stean, Marbella
1982, the now-retired English GM found a
fine refutation: h5 11.♕e5 f6! 12.♗xf6
d6 13.♕e4 ♗xg4! 14.♗xh8 ♔d7 15.♗d4
♗xh4+ and White resigned; it was not so
easy to prepare novelties in such sharp
positions, in the pre-computer era) 9...♕e7!
( Bronstein mentions that the variation
9...♗e7?? 10.♘f6+ ♔f8 11.♗h6#
was what first attracted him to the King's
Gambit ) . Now White can win the black
q u e e n w i t h 10.♘f6+ ♔d8 11.♗xc7+
♔xc7 12.♘d5+ , but after ♔d8 13.♘xe7
♗xe7 Black has too much material, as his
knight on h1 is not so easy to trap. ]
[ QUESTION: So what is the point of
Bronstein's 7.♘c3 ?
ANSWER: Well, as he puts it himself in
'David against Goliath', one day he asked
himself why White should drive away the
e n e my k n i g h t w i t h a p a w n ? W h y n o t
develop a piece at the same time? ]
[ Bronstein gives 7...d5 as safer, which is
certainly is, but the text leads to
fascinating complications, so we can be
grateful to Mr. Ostrup for entering into
the spirit of things. ]
♗g7 Bronstein criticizes this, but
[ H e r e c o m m e n d s i n s t e a d 8...♗d6!
, giving the striking variation 9.♕f3 ♘xh1
10.d4 c6 ( however, the computer pours a
10...♕xh4+ 11.♔d1 f5! 12.♘gf6+ ♔f7
and White can objectively resign) 11.♗xf4
♗xf4 12.♕xf4 cxd5 13.♘f6+ ♔e7
14.♕e3+! ♔f8 ( not 14...♔xf6? 15.♕g5+
with perpetual check. ]

Bronstein − Move by Move

 Bronstein − Move
31 by Move
[ 8...♗e7 is not mentioned by Bronstein, but
is also better for Black, according to the
computer.
QUESTION: So what was Bronstein
planning?
ANSWER: Well, one cannot know for sure,
but I suspect something similar to the
previous variation, such as 9.♕f3 ♘xh1
10.♕xf4 , although the silicon monster is
t h o r o u g h l y u n c o n v i n c e d , a f t e r ♗xh4+
11.♔d1 ♕g5! when the threat of 12...Qxg4
and 13...Nf2+ forces 12.♘xc7+ ♔d8
13.♕xg5+ ♗xg5 14.♘xa8 b6 and White's
knight faces a mu ch harder task than
Black's, when it comes to emerging from
the corner. ]
9.d4!? 0-0 QUESTION: Couldn't he have
taken the rook at once?
[ ANSWER: Probably. Bronstein describes
t h i s a s " d a n g e r o u s " a f t e r 9...♘xh1
10.♕e2+ ♔f8 11.♗xf4 ♕xh4+ 12.♔d2
d6 13.♘xc7 , but even here, the computer
sees no danger after ♗d7 .
QUESTION: So are we saying that this
whole game is just nonsense?
ANSWER: It is not necessarily so clear.
Computers are very bad at assessing
positions with long-range threats. They
look at such a position, count the pieces,
and then look for any threats which are
within their horizon. If they can't see any,
th ey assess the position accordingly.
Bronstein won many games against the
early generation of computers, by
accepting such positions, and just playing
on, gradually developing threats. Often,
his initiative would prove to be much
more dangerous than the computer realized,
and the assessment would suddenly change,
when it was already too late. Admittedly,
the modern generation of chess engines is
a great deal stronger than the ones
Bronstein − Move by Move
Bronstein faced, so it remains
questionable whether such an approach
could ever work with the modern
computers. Against a human opponent,
though, such attacks can be extremely hard
to defend, even if strictly speaking unsound.
Naturally, if one is afraid of risk, one
does not play 2 f4... ]
10.♗xf4 ♘xh1 11.♘h6+! Amazingly,
Bronstein claims that he used to analyse this
move, in his head, during lessons at school!
He waited over 50 years for a chance to play
it, which probably makes this the longest-
prepared opening variation in history.
QUESTION: So what is going on?
ANSWER: White now has quite a dangerous
attack, as even the computer acknowledges,
after a short think.
♔h8 QUESTION: Is this forced?
ANSWER: It seems not.
[ Bronstein's idea was 11...♗xh6 12.♕g4+
. No w aft e r ♔h8 ( 12...♗g7?? 13.♗h6
wins at once) , there follows 13.♗xh6 ♖g8
14.♗g5 with dangerous threats. Bronstein
t h e n g i v e s a s a s a m p l e l i n e ♕e8+
( t h e c o m p u t e r s h o w s t h a t 14...f6!
is still objectively better for Black:
15.♗xf6+ ♕xf6 16.♕xg8+ ♔xg8
17.♘xf6+ ♔f7 18.♘d5 ♘a6
, etc; however, such a line is not so easy for
Black to find at the board, and Ostrup's
decision to retain his defensive bishop is
entirely understandable) 15.♔d2 ♕e6
16.♕f4 and Black is defenceless: ♕xd5
17.♗f6+ ♖g7 18.♕h6 mates. ]
12.♕h5 ♕e8+ 13.♔d2 f5? The desire to
exchange queens is also very understandable,
but proves mistaken.
[ Bronstein himself, with computer help,
found the best defence, which is 13...♘f2!
. This forces an interesting draw, after
14.♖e1 ♘e4+ 15.♔d1 ( unfortunately,
 Bronstein − Move
32 by Move
15.♔c1? ♘f6 wins for Black, since the
rook on e1 is undefended) 15...♘f2+
16.♔d2 ♘e4+ with a repetition. Bronstein
adds that the attempt to avoid the draw with
17.♖xe4 ♕xe4 18.♗d3? loses, but his
variation is not quite correct. After ♕xg2+
19.♔c3 , he gives ♗xh6? ( instead, Black
should clear the g7-square with tempo in a
d i f f e r e n t w a y : 19...♗xd4+! 20.♔xd4
♘c6+ 21.♔c3 ♕g7+ when Black beats
off the attack and has a decisive material
advantage ) 20.♕xh6? ( but in this line,
20.♗xh6! wins for White ) 20...♕g7 . ]
14.♕xe8 ♖xe8 QUESTION: But once the
queens are off, surely White must be losing?
He is a whole rook down, after all!
ANSWER: Yes, but he still has a massive
lead in development, and the black king is
still not safe.
15.♗d3!?
[ Bronstein said it was impossible to
calculate the consequences of the variation
15.♘xc7 ♖e4! , b u t a f t e r 16.♖e1!
White is clearly on top, as the computer
shows: for example, ♖xf4 17.♖e8+ ♗f8
18.♗d3 ♔g7 19.♘xf5+ ♔f7 20.♖xc8
and it is all over. ]
15...♘a6 16.♖xh1 QUESTION: What on
earth is going on?
ANSWER: White has regained a piece, and is
now only an exchange down. He still has a
large development advantage, with all his
pieces active and in threatening positions,
while Black has his knight on h1 out of play,
his queenside undeveloped, and faces a
threat of Bxa6, followed by Nxc7. White has
more than enough compensation for his
material, as his attack continues.
c6?
[ Obvious, but 16...d6 was probably the
only chance as it does at least make a
small contribution to developing Black's
Bronstein − Move by Move
queenside, which cannot be said of the text
move. ]
17.♘f7+! ♔g8 18.♘d6 QUESTION: It
seems that Bronstein, almost as a matter of
principle, never moves an attacked piece.
ANSWER: Well, there was no doubt a sense
of impish pride in th at, but it is also a
logical consequence of the position. When
you are material down, but with the initiative,
it is vital to keep threatening the opponent
and not give him any respite. That often
means ignoring attacks on one's own pieces
and seeking ways to respond by creating
bigger threats oneself.
♖e6 19.♘e3 Now the knight finally moves,
but only after Black's rook has been lured to
an exposed square on e6. The black
queenside pieces are never likely to see the
light of day, and White's initiative has
become crushing.
♗h6
[ After 19...♗xd4 20.♘exf5 ♗xb2
the obvious
A) but the computer finds the far-from-
o b v i o u s , a n d e v e n s t r o n g e r 21.♖f1!
bringing the rook into the attack, and
threatening a deadly assault on the black
king. For example, b5 (stopping Bc4)
22.♘h6+ ♔g7 .
EXERCISE: Can you spot the lovely
forced mate?
ANSWER: 23.♗e5+!! ♔xh6 ( or
23...♗xe5 24.♖f7+ ♔xh6 25.♖xh7# )
24.♘f7+ ♔h5 25.♗e2+ ♔xh4 26.♖h1#
.
QUESTION: Beautiful! One does not
expect to see such mating attacks once
queens are off the board!
ANSWER: Maybe not, but with so many
other pieces on, the position cannot really
be described as an endgame. It is more
a queenless middlegame, and with a
 Bronstein − Move
33 by Move
massive development lead and an
ex p o s e d b l a ck k i n g , ma ti n g a t ta c k s
remain a feature of the play.;
B) 21.♗c4 wins for White, ]
20.♗xh6 ♖xd6 21.♗c4+ ♔h8 22.♘xf5
[ 22.♗g7+! ♔xg7 23.♘xf5+ was even
better, as Bronstein points out. ]
22...♖xh6
[ Bronstein shows that 22...♖g6 23.♖e1
♘c7 is a tougher defence, but then Fritz in
turns points out that 24.♗g5 ( Bronstein's
24.♗f4 is also good, but less convincing)
24...d5 25.♗d3 is still crushing for White:
for example, ♗xf5 ( or 25...♗e6 26.♘d6
♖gg8 27.♗f6+ ) 26.♗xf5 ♖xg5
( the rook has no good move: for instance,
26...♖d6 27.♗f4 ) 27.hxg5 , and all of
White's brilliance and inventiveness yields
an endgame with an extra pawn.
QUESTION: But the pawn is doubled, so is
White winning here?
ANSWER: Yes, I think so, given his much
more active pieces and the power of the
rook and bishop combination. There is no
g o o d w a y t o m e e t R e 7 , s i n c e i f ♖g8
28.♖e7 ♖g7 29.♖xg7 ♔xg7 30.♗c8 b6
31.♔e3 and White should win comfortably
enough. ]
23.♘xh6 d5 24.♗xa6
[ As in the last variation, the game comes
down to a winning endgame for White,
after 24.♗xa6 bxa6 25.♘f7+ ♔g7
26.♘e5 . He has an extra pawn (this time,
not even doubled), a dominant knight and
weaknesses among the enemy queenside
pawns. ]
1-0
Bronstein − Move by Move
17 C32
Chandler,C
Bronstein,D
Kent League (Game 8) 1996
[Steve Giddins]
This game is, it must be said, relatively one-
sided, after the opening, but I am pleased to
be able to include it, not least because it is
another hitherto unpublished Bronstein game.
The scene is a typical evening league match,
in the county of Kent, in south-east England.
In the back room of a pub, in the charmingly-
named city of Gravesend (so called because
it marked the end of the line of graves,
containing victims of the Great Plague of
London in 1665), a group of local
chessplayers gathered, after a hard day at the
office, for the customary evening thrash –
typically, 35 moves in 75 minutes, then a 15-
minute allegro finish.
On top board for the home team was local
accountant, Cliff Chandler, a strong (2300+)
amateur, former three-time British
Correspondence Champion, and a great lover
of gambits. When the visitors from Charlton
Chess Club turned up, their numbers included
a diminutive, balding Russian grandmaster,
who once drew a match against Botvinnik for
the world championship. Undaunted, and (I
hope!) with pint of real ale in hand, Cliff
trotted out his King's Gambit. The result was
as follows.
1.e4 e5 2.f4 d5 In view of Game 6, which
was one of the most important in
establishing the 'refutation' of the main line
of the Falkbeer, it is interesting that
Bronstein should choose the line as Black.
Alas, what he intended against his own
treatment remains unknown.
3.exd5 e4 4.d3 ♘f6 5.dxe4 ♘xe4 6.♗e3
This is an old line, which has always
 Bronstein − Move
34 by Move
remained a favourite with certain romantic
King's Gambit players, since it involves an
exchange sacrifice.
QUESTION: But doesn't it fail to meet the
threat of a check on h4?
[ ANSWER: That is the line that White is
hoping for. After 6.♗e3 ♕h4+ 7.g3 ♘xg3
8.♘f3 ♕h5 ( 8...♕e7 is a safer option for
Black, but after 9.hxg3 ♕xe3+ 10.♕e2
♕xe2+ 11.♗xe2 we reach an endgame,
with a similar structure to that in Game 6;
as we saw there, White has reason to
expect some edge here) 9.hxg3 ♕xh1
10.♕e2 White has a position in which he
has quite dangerous play for the
exchange. ]
[ After 6.♗e3 , instead of the rather dubious
check on h4, John Shaw recommended ♗d6
as the best defence, and this does indeed
offer Black good play. Instead of this,
Bronstein produced a move which is not in
the books. ]
6...♘d7!? QUESTION: So what is the idea?
ANSWER: The main point is to set up the
possibility of bringing this knight to f6, to
support his colleague. As we saw in the
Bronstein-Tal game, in many lines of the
Falkbeer, Black has problems with the
instability of his proud-looking steed in the
centre. If the knight can be maintained on e4,
it can be a very effective piece, but this is
often difficult to manage. Bronstein's move
addresses that concern.
7.♘d2!?
[ 7.♘f3 is perhaps the most natural, but
Chandler clearly preferred to dangle the
lure of a check on h4 in front of Black for
one move longer. However, after 7 Nf3,
White may well be better: for example,
♘df6 8.♗c4 ♗b4+ 9.c3 ♗d6 10.♘bd2
and it does not look as though Black has
enough for the pawn. ]
Bronstein − Move by Move
7...♘ef6!? QUESTION: That is funny! I
thought you said the idea of 6...Nd7 was to
maintain the knight on e4?
ANSWER: I did! But Bronstein's choice is
quite a clever reaction to White's last move,
since now it is harder for White to defend
the d5-pawn.
8.♘df3? After this, things turn against
White very rapidly.
[ 8.♗c4 leaves the bishop vulnerable to a
later ...Nb6, ]
[ whilst the non-developing move 8.c4?!
looks rather risky. Black has decent
compensation after ♗d6 and a quick ...0-0,
followed by ...Re8. ]
[ The computer recommends as best 8.♕f3
, defending the extra pawn and preparing
long castling. White may be a little better
after that, although the position is wholly
untested and it is difficult to draw any
categorical conclusions. ]
8...♗b4+! QUESTION: Why is this good?
ANSWER: Black develops with temp i,
exploiting the removal of the white queen's
knight to the kingside.
9.c3 ♕e7 This follow-up is the point of
Black's last move. White lacks a good way
to defend his bishop on e3.
10.♕e2
[ 10.cxb4 ♕xe3+ 11.♕e2 ♕xe2+
12.♘xe2 ♘xd5 is not particularly
att r act iv e, ei th er , b u t may have b een
slightly better than the text. Now Black
regains the pawn, with good play. ]
10...♘xd5! 11.♗d4? This leads by force to
an endgame, where White is just a pawn
down, with a broken position.
[ Relatively best was 11.♗d2 , although
Black is still for preference after ♗d6 . ]
11...♘xf4! 12.♕xe7+ ♗xe7 QUESTION:
White is just a pawn down, isn't he?
ANSWER: Yes, the opening has been a
 Bronstein − Move
35 by Move
disaster for him.
QUESTION: It seems strange that, just half
a dozen moves after Black's innovation,
White should already be almost busted!
ANSWER: Well, that is typical of the King's
Gambit, for both sides. It is an opening
which is very unforgiving, and where one
mistake can see the player's position
wrecked almost at once. This is very
attractive in some ways, and means that
King's Gambit players win a lot of very
quick and effortless games as White, but the
flipside of the coin is that, when it goes
wrong for White, he can suffer an equally
dismal fate.
13.♗xg7 Rather than settle for a pawn down
in a quiet, technical position, Chandler tries
to fish in murky waters, but, if you will
forgive the mixed metaphor, those who sup
with the Devil need a long spoon.
♖g8 14.♗h6
[ 14.g3 ♖xg7 15.gxf4 regains the pawn for
the time being, but after ♗d6 16.♘e2 ♘c5
17.0-0-0 ♗f5 Black has a wonderful
position, with two bishops, a weak pawn on
f4 and his pieces flooding into the holes
in the white position, especially d3.
Chandler has a different idea in mind –
he hopes to profit from the offside position
of the black knight if it captures on g2. ]
14...♘xg2+ 15.♔f2 ♘f6!
[ Even better than 15...♘h4 . The threat of a
check on g4 or e4 is highly unpleasant. ]
16.♗xg2 Sadly, there is nothing better.
[ The bishop on h6 has no decent retreat
square, as 16.♗c1 is obviously awful: for
example, ♘e4+ 17.♔e2 b6 with the very
embarrassing threat of ...Ba6+. ]
16...♘g4+ 17.♔f1 ♘xh6 18.♖e1 ♗e6
QUESTION: Hmm. White's thrashing about
does not seem to have improved his position.
ANSWER: Indeed. Black has a dream
Bronstein − Move by Move
position – an extra pawn, the two bishops,
and the white pieces shambolically huddled
in the corner of the board. The rest is of
course very easy for a grandmaster, but it is
still instructive to see how Bronstein keeps
the initiative and does not sit back on his
laurels and start playing routinely.
19.b3 0-0-0 20.♘d4 ♗h4 21.♖e4 ♖xg2!
A nice combination to wrap things up. Black
is sacrificing a piece, to break through with
his rooks for a decisive attack.
22.♖xh4
[ Forced, since 22.♔xg2 ♗d5 wins even
more material, ]
[ and 22.♘xe6 ♖f2+ is even worse. ]
22...♖xa2 23.♖xh6 ♗xb3! 24.♘ge2
[ There is no defence, as shown too by
24.h4 ♗c4+ 25.♔e1 c5 26.♘de2 ♗xe2
27.♘xe2 ♖a1+ , winning a rook. ]
24...♗c4 25.♔f2
[ If 25.♖g1 ♖xd4 26.cxd4 ♖xe2
and White cannot save his rook on h6 from
dropping to a discovered check, except by
27.♖g8+ ♖e8+ , which drops the other
rook instead. ]
25...c5 Diagram
(Diagram 28)
 Bronstein − Move
36 by Move
XHGFEDCBAY
1R+-+-+-+![
2zP-mKN+-+r" 28
3-+-+-zP-+#
4+-+-sNl+-$
5-+-+-zp-+%
6tR-+-+-+-&
7p+p+-+pzp'
8+-+-trk+-(
xhgfedcbay
Regaining the piece.
QUESTION: This is very nice play by
Bronstein, of course, but it seems surprising
that he should play such a sacrifice at move
21, when he could have won by simple and
prosaic means.
ANSWER: That was Bronstein all over. In
the first place, calculating the variations
resulting from 21...Rxg2! would not have
presented the slightest difficulty to him, even
at the age of 72 (as he was when this game
was played), and when playing at a fast time-
limit. And secondly, he was always a
restlessly creative player, who would prefer
the combinative path to the prosaic technical
exploitation of extra material.
26.♖c1!? A nice trick, hoping to create
tactical trouble with the pin on the soon-to-
be-opened c-file, but it is hopeless, of course,
like every other move for White.
cxd4 27.cxd4 b5 28.♔f3 Thus, White
saves the piece, but his position remains lost.
♔b7
[ 28...♖xe2 allows 29.♖xc4+! , although
Black is actually winning even this rook
ending. ]
Bronstein − Move by Move
29.♘c3 ♖d2 30.♖xh7 ♖8xd4 31.♘xb5
♖2d3+
[ 31...♗xb5 32.♖xf7+ ♗d7 is a simple win
anyway, but Bronstein deprives his
opponent even of the one check. ]
32.♔f2
[ 32.♔g2 ♖g4+ forces the king to the f-file
anyway. ]
32...♖f4+ 33.♔e2 ♗xb5 A highly
entertaining, if lightweight game. I am
indebted to Cliff Chandler, for kindly
supplying me with the game score.
0-1
18 B10
Tal,M
Bronstein,D
Exhibition match, Moscow (Game 9) 1966
[Steve Giddins]
This is the only game in the book that was
not played in a competitive tournament or
match. It was part of an exhibition, given by
these two great masters, in 1966. As might
be expected, with the responsibility of prizes
and rating points taken away, the game turns
into a light-hearted and highly entertaining
affair, which is well worth rescuing from its
obscurity.
1.e4 c6 2.d3 The King's Indian Attack is
not generally thought of as the harbinger of a
tactical slugfest, but it does have the merit
of avoiding a fair bit of established theory,
and the players soon manage to turn it into
something highly original.
e5 3.f4 d5 QUESTION: Looks more like a
King's Gambit!
[ ANSWER: Indeed. If White were now to
meet 3...d5 with 4.exd5 , we would reach
a position from Nimzowitsch's counter-
gambit 1 e4 e5 2 f4 d5 3 exd5 c6 4 d3 (4
 Bronstein − Move
37 by Move
Nc3 is regarded as best). Instead, Tal
decides to get a piece out at last. ]
4.♘f3 dxe4 5.♘xe5 ♕h4+?!
[ In the spirit of 19th century fun, but
objectively, 5...♘f6 is stronger, with equal
chances. ]
6.g3 ♕e7 7.d4
[ This disposes of the threats along the e-file,
b u t i n f a c t , 7.♘c3 could also have been
played. Then the attempted larceny with
exd3 8.♗xd3 f6? fails to 9.♕h5+ ♔d8
( 9...g6 10.♗xg6+ hxg6 11.♕xh8
is also clearly better for White) 10.0-0
with a ferocious attack for the piece. ]
7...♘h6 8.♗c4 ♗e6 9.d5? This leads to
great entertainment, but is objectively
dubious. White is not sufficiently ahead in
development to justify this early opening of
the position.
cxd5 10.♗xd5 f6 11.♘c4
[ 11.♕h5+ g6 12.♘xg6 hxg6 13.♕xg6+
♔d8 does not give White enough. ]
11...♘c6 12.♘e3 Now White comes under
pressure down the d-file, but it was difficult
to suggest anything better.
QUESTION: Why can't he just exchange on
e6?
[ ANSWER: After 12.♗xe6 ♕xe6 13.♘e3
♖d8 White's position is very unpleasant –
he is behind in development, his queen is
exposed and he has also contracted some
nasty weaknesses on the kingside light
squares. ]
12...♖d8
[ 12...0-0-0! looks even stronger. ]
13.c4 ♕f7 14.♘c3 ♗c5 Offering a pawn to
open more central lines.
15.♘xe4 QUESTION: That looks very
risky!
ANSWER: Yes, I am not sure Tal would
have grabbed this pawn in a proper game.
But it is clear that the present encounter is
Bronstein − Move by Move
being played in a 19th century spirit, a time
when gambit pawns were accepted, almost as
a matter of honour.
♗d4?!
[ 15...♗xe3 16.♗xe3 0-0 looks very strong,
as after 17.0-0 ♗xd5 18.cxd5 ♖fe8
White's minor pieces are hanging
awkwardly on the e-file and the extra pawn
will soon fall. ]
16.f5 ♗xd5 17.♘xd5 ♘xf5 QUESTION:
This looks like a disaster for White! His king
is caught in the centre.
ANSWER: Yes, but I think Tal was counting
on the strength of the coming counterattack.
18.♕g4 ♘d6
[ 18...♘ce7 was also good, but Bronstein
enters into th e spirit o f th e game, b y
inviting the white queen into his parlour. ]
19.♘xd6+ ♖xd6 20.♕c8+ ♖d8 21.♘c7+
♔e7 22.♕xb7 I think this is probably what
Tal had seen when he played his 16th move.
QUESTION: Black's king looks very
ex posed! Surely it is now he who is in
trouble?
ANSWER: It turns out that he is not, as
White has too few pieces with which to
attack. In fact, it is the white king which
will soon come under a crushing attack.
♕xc4 23.♗f4 Logically bringing up another
piece.
[ There is nothing to be had from the
discovered checks: for example, 23.♘d5+
♔f8 24.♘f4 ♗b6 and White is
defenceless. ]
23...♘e5
[ 23...♕c2! was even stronger, but the text
is good enough. Besides, we should not
criticize the play too severely. It is
obvious that entertaining the crowd was
the main purpose, rather than ruthless
pursuit of the best moves. ]
24.♖c1 QUESTION: Why doesn't he
 Bronstein − Move
38 by Move

eliminate the knight, while he has the

chance?
XHGFEDCBAY
[ ANSWER: Alas, this allows a forced mate, 1-+-+-tR-+!
starting with 24.♗xe5 ♗f2+! 25.♔xf2 2zP-+-mK-zPP" 29
♖d2+ and mate in five. I will leave you,
dear reader, to work out the details. ] 3-zP-+-+-+#
24...♘d3+ 25.♔d2 ♘xc1 A remarkable
position to have arisen in a game between an
4+-vL-vlq+-$
ex-world champion and an ex-challenger! 5-+-+k+-+%
QUESTION: It looks like a total mess!
ANSWER: Actually, White is completely 6+-zp-+-+-&
lost. He has too few pieces with which to 7pzp-wQ-+-zp'
attack the black king successfully, whereas
his own monarch is facing imminent 8tr-+-tr-+-({
extinction.
26.♘d5+ ♔e6! QUESTION: Playing to the
xhgfedcbay
gallery again? ♗e3+! A delightful finish. Black clears the d-
[ ANSWER: Absolutely not! The move is file with tempo, putting his bishop en prise
forced, since 26...♔f8?? allows mate after to no fewer than three white pieces.
27.♕e7+ ♔g8 28.♕e6+ ♔f8 29.♗d6+ [ Strictly speaking, a quiet move such as
♖xd6 30.♕e7+ ♔g8 31.♕e8# 28...♕b5 was stronger, but who cares about
. Admittedly, in the circumstances, that? ]
Bronstein would probably have played 26... 29.♗xe3
Ke6, even if it had not been forced. ] [ 29.♔xe3 ♕xc1+ is obviously hopeless, ]
27.♕e7+ ♔xd5 [ whilst 29.♕xe3 ♖he8 30.♕d3+ ♕xd3+
[ 27...♔f5 is another solution: 28.g4+ 31.♔xd3 leaves White with a lost endgame.
♔xg4 29.♕e6+ f5 and the checks run Objectively, this last variation is what he
out. ] should have played, as it allows the
28.♖xc1 Diagram longest resistance, but of course, it was
tacitly understood by both players that
(Diagram 29) this game should not end with a 'boring'
endgame. ]
29...♔c6+ 30.♔e1 Diagram

(Diagram 30)

Bronstein − Move by Move

 Bronstein − Move
39 by Move

XHGFEDCBAY ANSWER: Yes, it is not one of the main

lines here.
1-+-mK-tR-+! [ There is a great deal of modern theory on
2zP-+-+-zPP" 30 the line 5.♗d3 ♘gf6 6.♘g5 , with all
manner of traps associated with the
3-zP-vL-+-+# sacrifice Ne6 in various lines, but none of
this was known at the time of the present
4+-+-+q+-$ game. Besides, even had it been, I doubt
5-+-+-+-+% that Bronstein would have been terribly
interested in going down such lines. He
6+-zp-+k+-& preferred to create a less-explored position
7pzp-wQ-+-zp' on the board. ]
5...♘gf6
8tr-+-tr-+-({ [ 5...e5 is the obvious response to White's
xhgfedcbay last, but Bronstein rejects this on the
grounds that after 6.♘f3 exd4 7.♘xd4
♕xc1+! "the knights in the centre would be too
[ 30...♕xc1+ 31.♗xc1 ♖he8 forces an strong". ( White does, indeed, have some
easily winning ending. A very initiative here, but the moves 7.♕xd4;
lightweight gamelet, but highly and even 7.♗g5 both look even better. )]
en tertaining for the spectators, and a 6.♘g3 QUESTION: Why lose a tempo
striking example of what consenting retreating the knight?
Soviet Grandmasters could get up to, even ANSWER: Bronstein explains that he did not
in public. ] want to exchange on f6, which would free
0-1 the black bishop on c8. In general, White has
more space in such Caro positions, thanks to
his pawn on d4, so exchanges are unlikely to
19 B17 be in his favour.
Bronstein,D e6 7.♘f3 ♗e7
Zamikhovsky,A [ 7...♗d6 is a little more active. The h2-b8
Leningrad (Game 10) 1970 diagonal is frequently an important one in
[Steve Giddins] such structures, and a set-up with ...Bd6
and ...Qc7 generally offers Black better
1.e4 c6 2.d4 d5 3.♘c3 dxe4 4.♘xe4 ♘d7 prospects than having the bishop passive on
The move generally associated with Smyslov, e7. ]
although Flohr played it as early as the 8.♗d3 c5 9.0-0 0-0 10.♘e5 QUESTION:
1930s. How do you assess the position?
[ Bronstein himself frequently played the ANSWER: White is a little better. He has
Caro-Kann as Black, but generally more space and his development is freer, but
preferred the more pugnacious variation Black is very solid.
4...♘f6 5.♘xf6+ gxf6 . ] ♕c7 QUESTION: How come Black did not
5.c3 QUESTION: That looks rather quiet! isolate the white d-pawn when he had the

Bronstein − Move by Move

 Bronstein − Move
40 by Move
chance?
[ ANSWER: As becomes clear, White does
not object to such an IQP structure. After
10...cxd4 11.cxd4 we have a typical
scenario, where the IQP gives White
greater central space and an outpost on e5,
whilst Black can play to occupy the outpost
on d5, in front of the pawn. Black avoids
this, mainly because he wants to develop
his queen's bishop actively on the long
diagonal, for which purpose he needs to
guard the c6-square, to stop ...b6 being met
by Nc6. That in turn means he needs to put
his queen on c7, and therefore does not
wish to open the c-file. ]
11.♕e2 b6 12.f4 ♖e8? QUESTION: Why is
that wrong? It looks very natural to me.
ANSWER: Actually, one frequently sees this
type of move in these structures, but it is not
always good. Black was obviously disturbed
by the prospect of f4-f5, which White's last
move set up, and so he hurries to put his
rook opposite the white queen, and to free f8
for his knight.
QUESTION: So what is wrong with that?
ANSWER: The problem is the weakening of
the f7-square, which White quickly exploits,
and that is often the problem with moving
the rook from f8 in such positions.
[ Black should instead prefer the logical
12...♗b7 , after which 13.f5 can be met by
cxd4 14.cxd4 ♕d6! when White has some
problems defending his d4-pawn
satisfactorily. ]
13.♘e4! All of a sudden, some nasty threats
appear over the head of the black king, all
thanks to the weakness of f7. 14 Ng5 is one
threat, whilst White also has the possibility
of the more direct 14 Nxf7! Kxf7 15 Ng5+
and disaster on e6.
h6
[ QUESTION: So maybe Black should play
Bronstein − Move by Move
13...♘xe5 , getting rid of the powerful
knight?
ANSWER: Unfortunately, he cannot do
this, as Bronstein points out, because of
14.♘xf6+! ♗xf6 15.fxe5 ♗e7 and now
16.♕e4 w i n s a r o o k , ( whilst my little
German friend informs me that 16.♗xh7+!
is forced mate in 11: ♔xh7 17.♕h5+ ♔g8
18.♕xf7+ ♔h7 19.♖f3 and desperadoes
on h4 only delay the inevitable. Once again,
in this line, we see the weakness of f7. )]
[ QUESTION: So the text move was
forced?
ANSWER: Not quite. Bronstein
r e c o m m e n d s 13...♘xe4 14.♕xe4 ♘f6
15.♕xa8 ♗b7 16.♕xe8+ ( but 16.♕xa7
♖a8 17.♕xa8+ ♗xa8 wins an extra pawn
by comparison with his line, and with two
rooks and a pawn for the queen, White
m u s t b e b e t t e r h e r e) 16...♘xe8
"and Black has nothing to fear". ]
[ It may be that the best move was 13...♖f8
, but it takes a character with the strength
of a Botvinnik or a Karpov to play such a
move, acknowledging the error of the move
before. ]
QUESTION: But what is wrong with
Zamikhovsky's move? It looks natural,
preventing the other white knight coming to
g5.
ANSWER: Yes, but there are two problems.
Firstly, it weakens the kingside severely,
especially g6, and secondly, and even more
to the point, it does not actually prevent the
knight coming to g5.
14.♘g5! Told you so!
♖f8 QUESTION: But what was Bronstein
going to do if the knight was taken?
[ ANSWER: Actually, in his notes in 'The
Sorcerer's Apprentice', he does not give
any variations at all, merely stating that
"Taking the knight is too dangerous". But
 Bronstein − Move
41 by Move
in view of what happens in the game, it
seems that Black might as well try this,
although deep analysis shows that it still
loses. 14...hxg5 15.fxg5 and now Black
has a choice of evils:
A) 15...♘d5 16.♕h5 forces mate in
four,;
B) w h i l s t a f t e r 15...♘xe5 16.dxe5
, Black also faces serious problems. The
only way to stay on the board for the
time being is
B1) 16...♘d5? again allows a forced
mate after 17.♗h7+! with such lines as
♔xh7 ( if 17...♔f8 18.♖xf7+! ♔xf7
19.♕h5+ ♔f8 20.♗g6 ♘f6 21.gxf6;
or 17...♔h8 18.♕h5 and only various
desperate piece sacrifices on g5, f6 and
f4 can delay the mate) 18.♕h5+ ♔g8
19.♕xf7+ ♔h7 20.g6+ ♔h8
21.♕xe8+;
B2) 16...c4 17.♗c2 ♗c5+ 18.♔h1
♘d7 , but here simply 19.♕e4 g6
20.♕xa8 wins, leaving White an
exchange ahead, and his queen is not
getting trapped, since the undefended
rook on e8 means that Black cannot
play ♗b7 .;
C) QUESTION: So what should Black
do?
ANSWER: The best (only) defence is to
return the piece with 15...cxd4 16.cxd4
♗b7 17.gxf6 ♗xf6 and hope to defend,
but now White again wins with the
t h e m a t i c s a c r i f i c e 18.♗h7+! ♔f8
(relatively best, as other moves run into
19 Qh5) 19.♗f4 and Black is just getting
mauled, without even having any extra
material to show for his pains.;
D) 15...-- ]
So, taking on g5 is not possible. After the
text, Black looks again to have repulsed the
threats. With f7 defended, he threatens to
Bronstein − Move by Move
take on g5, but now a fresh blow greets him.
15.♘g6! ♖e8 QUESTION: What happens if
we take on g6 now?
[ ANSWER: Bronstein actually says this is
best, although after 15...fxg6 16.♕xe6+
♔h8 17.♕xe7 Black is just losing: 18 Ne6
is a terrible threat, and hxg5 ( 17...♕d8
18.♘f7+ wins the queen) 18.fxg5 ♘h7
19.♖xf8+ ♘hxf8 20.♕f7! leaves Black
defenceless against the threat of Bc4: ♗b7
21.♗c4 ♘h7 22.♗f4 ♕d8 23.♖e1
and Re7 will decide. ]
16.♘xe6! QUESTION: This is amazing!
[ ANSWER: Yes, it is delightful the way
Bronstein keeps sacrificing his knights,
move after move. This is the third straight
move that he has offered one knight or the
other, and now he self-forks them both!
B u t n e i t h e r c a n b e c a p t u r e d : 16.♘xe6
fxe6? ( while 16...fxg6? just leaves the
queen en prise to 17.♘xc7 ) 17.♕xe6+
♔h7 18.♘xe7+ mates quickly. ]
16...♕d6 Yet again, Black finds a way to
avoid taking the knights, and yet again,
White seems to have run up against a brick
wall. How does he extricate his kamikaze
cavalry?
17.f5!! "Banzai!". This stunning move
simply frees the f4-square for White's bishop,
creating the threat of Bf4, followed by Nc7.
There is no defence.
QUESTION: Presumably Bronstein had to
have seen this when entering the whole
sequence starting with 14 Ng5, didn't he?
ANSWER: Yes, he must have done.
fxg6
[ There is nothing better: 17...fxe6 18.♗f4
♕c6 19.♗b5 ♕b7 20.♕xe6+ ♔h7
21.♗c6 wins everything. ]
18.♗f4 ♕c6 19.♗b5 ♕e4
[ 19...♕b7 20.♘c7 wins, ]
[ whilst the desperado deflection sacrifice
 Bronstein − Move
42 by Move
19...♗a6 does not really help after simply
20.♗xa6 , when Nc7 is still impossible to
prevent, as ♗d6? loses material to 21.♗b5
.]
20.♕xe4 ♘xe4 21.fxg6!? QUESTION: Eh?
[ ANSWER: Indeed! A normal player
would have played 21.♘c7 ]
[ or 21.♗c6 , simply annexing a large lump
or two, but then again, a normal player
would never have reached this position in
the first place. Bronstein sees that Black
cannot meet all these threats anyway, so
he rather impudently takes the pawn on g6,
and invites his opponent to contemplate his
fate. ]
21...♘df6
[ Black cannot meet all the threats. For
example, if 21...♗d6 , then just 22.♗xd6
♘xd6 23.♘c7 ♖d8 24.♗xd7 ♗xd7
25.♘xa8 ♖xa8 26.dxc5 , and if bxc5?
( 26...♘e4 is more tenacious, of course, but
W h i t e i s s t i l l w i n n i n g a f t e r 27.♖ae1
, for example) 27.♖ad1 wins even more
material. ]
22.♘c7 ♗d7 23.♗xd7 ♘xd7 24.♖ae1
Bronstein notes with satisfaction that "the
last white piece comes into action."
♘df6 25.♘xa8 ♖xa8 26.♖xe4! ♘xe4
27.♖e1 White wins two pieces for the rook,
eventually reaching an ending with several
extra pawns.
♗d6 28.♖xe4 ♗xf4 29.♖xf4 It is all over
and Black could resign with a clear
conscience.
♖e8 30.dxc5 bxc5 31.♖f7 ♖e6 32.♖xa7
♖xg6 33.♖c7 ♖d6 34.b3 ♖d2 35.♖xc5
Another superb combinational display by
Bronstein. His handling of his "mounted
regiment" was beautiful.
1-0
Bronstein − Move by Move
20 B12
Bronstein,D
Hunt,H
Maidstone (Game 11) 1994
[Steve Giddins]
1.e4 c6 2.d4 d5 3.f3 QUESTION: That is a
pretty clumsy-looking way to meet the Caro-
Kann!
ANSWER: Maybe, but in recent years, the
so-called Fantasy Variation has enjoyed
something of a resurgence – or, given that it
was never popular in the first place, perhaps
that should be 'surgence'? It is a sharp and
interesting way to play, if White is willing to
gambit a pawn. Anyone interested in the line
should undoubtedly check out Alexey
Bezgodov's outstanding 2013 book on the
line, 'The Extreme Caro-Kann'.
e6 The most solid response.
[ The sharpest choice is 3...dxe4 4.fxe4 e5
(threatening 5...Qh4+) 5.♘f3 when exd4?!
( 5...♗g4 is more prudent, though still quite
sharp ) 6.♗c4 is rather dangerous for
Black. ]
4.♘c3 ♗b4 5.♘ge2!? An interesting and
rare move.
[ Bezgodov recommends the pawn sacrifice
5.a3 , ]
[ whilst the older move 5.♗f4 is also very
respectable. ]
QUESTION: This is a strange position!
ANSWER: Yes. It looks quite like the 4 Ne2
variation of the French Winawer, but in that
line, nobody has ever thought of following
up with 4...c6 5 f3...
5...dxe4 6.fxe4 ♕h4+ QUESTION: Is this
good?
ANSWER: I think it is probably OK.
[ The move 6...e5 was recommended in the
'British Chess Magazine' at the time, as
being more solid, but 7.♕d3 followed by
 Bronstein − Move
43 by Move
Be3 and long castling gives White
reasonable chances. ]
7.♘g3 Revealing the point of White's 5th
move. The e4-pawn is defended.
♘f6 8.e5 ♘e4!?
[ 8...♘d5 is more solid, when Black should
not have any problems. Instead, the text
instigates a long and sharp variation. ]
9.♕d3 ♘xg3 10.♕xg3 ♕xd4 QUESTION:
Is this right? Exchanging queens looks safer.
ANSWER: Having said 'A' in the form of 8...
Ne4, I think Black should say 'B'.
[ After 10...♕xg3+ 11.hxg3 White has
some advantage in the ending. The black
light-squared bishop is lamentably passive
and White has far more space.
QUESTION: But what about his doubled
pawns?
ANSWER: I don't think that is a
significant factor. The pawns are not on
an open file, and so are not really weak,
whilst the open h-file can be useful for
White's rook. ]
11.♕xg7 ♗xc3+ 12.bxc3 ♕xc3+ 13.♔f2
QUESTION: Crikey! Not exactly the sort of
position one associates with the Caro-Kann,
is it?
ANSWER: Indeed not, and that was probably
part of the attraction for Bronstein.
QUESTION: What on earth is going on?
ANSWER: Disappointingly, the position is
probably just equal, if one believes the
computer. There is a distinct chance that
some sort of perpetual will result. However,
any misstep by either side can be
immediately fatal.
♖f8 QUESTION: Why not take a pawn with
check?
[ ANSWER: There seems no objective
reason not to. There could follow
13...♕xc2+ 14.♗e2 ♕c5+ 15.♗e3 ♕f8
16.♕g4 and White has compensation for
Bronstein − Move by Move
the pawn, in a thoroughly unclear position.
The dark squares in Black's camp are very
weak, but so is the e5-pawn. ]
14.♗h6 ♘d7 15.♗d3 f6? This turns out to
be a serious mistake.
[ Black should have accepted an endgame,
an exchange down but with two pawns,
after 15...♕xe5 16.♕xe5 ♘xe5 17.♗xf8
♔xf8 . Chances would be about equal. The
text exposes the black king far too much. ]
16.♖ae1!?
[ T h i s i s v e r y s t r o n g , b u t 16.♕xh7!
looks even better, as 17 Bg6+ is a terrible
threat. For example, ♕d4+ 17.♔g3!
♕xe5+ 18.♗f4 ♕c3 19.♖ae1
and the black king is dreadfully vulnerable,
whilst she has no obvious way to complete
her development. ]
16...b6
[ 16...fxe5+ 17.♔g3 does not help at all.
White will continue with Qxh7 and Bg6+,
with a lethal attack. Hunt's move attempts
to develop the queenside, in the hope of
being able to evacuate the black king by
means of long castling, but she never gets
the chance. ]
17.♕xh7 ♕d4+
[ The computer shows the line 17...♘xe5
18.♖xe5 fxe5+ 19.♗xf8 ♕d2+ 20.♔f3
♔xf8 and now simply 21.♖f1 and Black is
defenceless against a discovered check. ]
18.♔g3 ♘xe5 19.♖xe5!?
[ Bronstein could not resist this sacrifice,
which is still very good for White, but
objectively, he should play the simple
19.♗g6+ ♘xg6 ( if 19...♔d8 20.♖d1 )
20.♕xg6+ ♔e7 21.♗xf8+ ♔xf8 22.♖hf1
, winning.
QUESTION: Surely Bronstein could see
that, couldn't he?
ANSWER: Yes, of course, but he simply
could not resist the more spectacular line.
 Bronstein − Move
44 by Move
He had doubtless seen his 21st move, and
once seen, such a move is hard to refrain
from, especially for such a romantic as
Bronstein. ]
19...♕xe5+ 20.♗f4 ♕c3 White needs to
unpin his bishop, after which the attack will
be crushing. Therefore...
21.♔h4!! It is disappointing that this should
be the computer's first choice, after just a
mi l l i s e c o n d o f t h o u g h t . B l a c k i s q u i t e
defenceless.
f5 Shutting out the bishop from g6, but now
it simply reroutes to another diagonal.
22.♗e2! ♕f6+ 23.♔h3 Steinitz would
doubtless have loved the way in which the
white king is able to defend itself in this way.
♖f7
[ QUESTION: But wait a minute, isn't
23...♖h8 trapping the queen?
ANSWER: No! White continues simply
24.♗h5+ unpinning, and forcing mate in
two: ♕g6 25.♗xg6+ ♔d8 26.♗c7# . ]
24.♗h5? A blunder, which could have cost
Bronstein the fruits of his play.
[ There was a simple win with 24.♕g8+
♖f8 25.♗h5+ ♔e7 26.♕h7+ ♖f7 27.♗xf7
♕xf7 28.♗d6+ ♔e8 29.♕h8+ ♔d7
30.♖d1 , etc. ]
24...♔e7? Returning the compliment.
[ After 24...♗d7! 25.♖d1 0-0-0 26.♕xf7
♕xf7 27.♗xf7 it looks as though White is
winning easily, with his extra piece, but
the eagle-eyed computer points out that the
bishop on f7 is not escaping after ♖h8+
28.♔g3 ♖h7! . The position is only a draw
after 29.♗xe6 ♗xe6 .
QUESTION: Amazing!
ANSWER: Yes, computers are desperately
tricky in such positions. I am sure neither
player even suspected the existence of such
a fiendish trick, and one cannot blame
them. ]
Bronstein − Move by Move
25.♗xf7 ♕xf7 26.♗d6+ ♔f6
[ Or 26...♔e8 27.♕h8+ ♔d7 28.♖d1
and there is no defence: for example, ♗b7
29.♗b8+ ♔e7 30.♕d8# . ]
27.♕h6+ ♕g6 28.♗e7+
[ 28.♗e7+ ♔f7 29.♕f8# is mate, and
otherwise the queen is lost.
QUESTION: Presumably, Bronstein would
have been choked to discover the missed
opportunity at move 24?
ANSWER: Actually, I am not sure he
would. Yes, it spoils the integrity of the
finish, but the remarkable defensive
resource 27...Rh8+ 28 Kg3 Rh7, trapping
the bishop on f7, would have delighted
Bronstein, I think. He was very much an
artist, who loved the richness of the
resources on the chessboard, so any
disappointment at the flaw in his game
would have been comp ensated for by
seeing such a lovely defensive trick. ]
1-0
21 C13
Kaplan,J
Bronstein,D
Hastings (Game 12) 1975
[Steve Giddins]
1.e4 e6 2.d4 d5 3.♘c3 ♘f6 4.♗g5 dxe4
5.♘xe4 ♗e7 6.♗xf6 gxf6 QUESTION:
That is a surprise? Can this really be good?
ANSWER: It looks odd at first sight, but it is
a reasonable option that has been known
about for a century or more. By recapturing
with the pawn, Black strengthens his central
control and also has the possibility of using
the half-open g-file. It has obvious
similarities with the Caro-Kann line 1 e4 c6
2 d4 d5 3 Nc3 dxe4 4 Nxe4 Nf6 5 Nxf6+
gxf6, which was also a favourite of
 Bronstein − Move
45 by Move
Bronstein's.
7.♘f3 ♘d7!? This was a favourite move of
Bronstein's in this position.
[ The more common treatment is 7...b6
, by means of which Black aims for a
restrained development, along the lines of
...Bb7, ...c6, ...Qc7, ...Nd7 and ...0-0-0.
This somewhat Hedgehog-like set-up is
very solid and offers Black reasonable
chances. It has been played a number of
times by Ulf Andersson. ]
[ The line with 5...gxf6 experienced a
revival of popularity about 20 years ago,
when the restlessly creative Alexander
Morozevich essayed it a number of times,
with a different interpretation, starting with
7...a6 . In this treatment, Black plays an
early ...b5, with an extended fianchetto.
Morozevich won several nice games with
the line. It is no coincidence that he and
Bronstein should both be attracted to 5...
gxf6, as they were similarly creative in
their approach to the game. ]
8.♕d2
[ An old Bronstein experience with the same
line saw his opponent try to break up the
b l a c k p a w n f r o n t w i t h t h e i mmed i a t e
advance 8.d5 . However, this failed to
yield anything, and after the sequence ♘b6
9.♗b5+ ♗d7 10.♗xd7+ ♕xd7 11.dxe6
♕xe6 12.♕e2 0-0-0 13.0-0 f5 14.♘g3
♕xe2 15.♘xe2 ♗f6 Bronstein later
claimed Black is already slightly better in
the endgame, and he went on to win in
Averbakh-Bronstein, USSR Championship
Semi-final, Leningrad 1946.
QUESTION: But why should Black be
better here? His pawns are doubled.
ANSWER: Yes, but they are not on open
files, so their vulnerability is limited,
whilst Black's bishop is the strongest
minor piece on the board, in this open
Bronstein − Move by Move
position. ]
8...c5 Annotating the game in the January
1976 issue of 'CHESS', Bronstein confessed
that he was slightly nervous about playing
this move, but felt that "it had to be done".
9.d5
[ White should probably have played
9.0-0-0 when Bronstein intended f5
10.♘c3 cxd4 11.♕xd4 ♗f6 followed by
castling short.
QUESTION: Again, this looks pretty shaky
to me. Isn't short castling dangerous, with
the g-file open?
ANSWER: Even more to the point, Black
cannot get his king castled in this line after
12.♕a4! , -- ( since 12...0-0? loses a piece
after 13.♗b5 .; It seems to me that Black is
i n s o m e t r o u b l e h e r e , a s e v e n 12...a6
does not prevent 13.♗b5 , when White's
position still looks threatening after ♖b8
14.♗xd7+ ♗xd7 15.♕f4 ♖c8 16.♖he1
, etc; ♗xc3 17.bxc3 ♕c7 18.♕d4
does not solve Black's problems. )]
9...f5 10.dxe6? Bronstein records that this
was played after long thought, but it seems
to be a clear mistake, after which Black
escapes from his problems and even assumes
the better of things.
[ Instead, 10.♘c3 keeps up the pressure,
when Black's position is still rather
unpleasant.
QUESTION: It doesn't look as though
Bronstein's opening was up to much! Does
this mean that 5...gxf6 is no good?
ANSWER: Not necessarily. I think
Bronstein's follow-up with 7...Nd7 is
probably inaccurate, fond though he
obviously was of the move. In general,
Black has to be quite careful in the early
stages of these positions, as if he breaks
out prematurely with moves such as ..c5
and ...f5, his lack of development and
 Bronstein − Move
46 by Move
uncastled king can rebound on him. I
don't think Black has any serious problems
in the Andersson line with 7...b6, but such
a rather restrained and passive set-up is
not to everybody's taste. ]
10...fxe4 11.exd7+
[ 11.exf7+ was the alternative. Bronstein
reports that his opponent rejected it because
of ♔xf7 12.♕f4+ ♘f6 13.♘e5+ ♔e8
14.♗c4 ♖f8 . However, Bronstein did not
like this at all.
QUESTION: So what do you think? ]
[ ANSWER: Well, Bronstein's instinct was
correct, even if not his variations. In the
first place, 11.exf7+ ♔xf7 12.♕f4+ ♘f6
13.♘e5+ ♔e8 does in fact lose by force –
i n s t e a d o f 14.♗c4? , ( White just plays
14.♗b5+ and Black can resign at once. )]
[ After 11.exf7+ ♔xf7 12.♕f4+
Bronstein's intended ♔g7 is therefore the
correct move, when he thought he was
allowing perpetual check, but the computer
points out that this is not so. Black is just
much better after 13.♕g4+ ♔f8 14.♕f4+
♗f6 and the checks run out. ]
11...♕xd7 12.♕c3 This is the critical
position. The other reason Kaplan did not
take on f7 last move is that he assessed this
position as being good for him.
QUESTION: Well, surely it is, isn't it?
ANSWER: It looks that way, as the black
rook on h8 is hanging, White is ahead in
development, and he has the better pawn
structure. However, things are not so simple,
as Black proves able to untangle successfully,
after which his two bishops prove very
strong.
0-0 QUESTION: Is the king really safe here?
It looks exposed.
ANSWER: The point is that White does not
have the pieces with which to harass the
enemy king. The queen looks very strong on
Bronstein − Move by Move
the long diagonal, but White has no other
pieces that can support the attack, whilst
Black's dark-squared bishop does an
excellent job of protecting the dark squares
around his king.
13.♘d2 ♕f5 14.0-0-0 ♕xf2 15.♘xe4?
Bronstein identifies this as the decisive
mistake As he puts it, "White ties himself up
regaining the pawn".
[ He recommends 15.♗c4 (at least, I think
that is what he recommends – 'CHESS',
which was in descriptive notation in those
days, actually gives the illegal move 15 B-
N4) "with equality", an assessment with
which the computer agrees. ]
15...♕f4+ Forcing the knight back into a self-
pin. From now on, Black's pieces flow into
the game with threats at almost every move.
16.♘d2 ♗g4 17.♖e1 ♗g5 The pin is deadly,
as White is unable to evacuate his king to b1,
whilst Black threatens to bring a rook to the
d-file.
18.♗d3 ♖ae8 19.♖ef1 QUESTION: Does
White have anything better?
[ ANSWER: Sadly not. For example,
19.♖xe8 ♖xe8 20.h3 and now simply ♗c8!
followed by ...Qf2, the point being that
21.♖f1? allows mate in two after ♕xf1+
22.♗xf1 ♖e1# .
QUESTION: It is shocking how quickly
the position has turned against White. How
did this happen?
ANSWER: It is all the consequence of his
15th move, which landed him in a fatal
pin. ]
19...♕e3 20.h3 ♗e2 21.♖f5
[ 21.♗xe2 ♕xc3 22.bxc3 ♖xe2 23.♖d1
♖d8 nets a piece at once. ]
[ Q U E S T I O N : B u t w h a t a b o u t 21.♖e1
, pinning Black's own bishop?
A N S W E R : T h e n t h e r e f o l l o w s ♕f2
when White is still tied hand and foot. The
 Bronstein − Move
47 by Move

threat is 22...Bxd3, so there is no choice Alekhine's Defence.

b u t t o e x c h a n g e o n e 2 . A f t e r 22.♗xe2
♖xe2 23.♖xe2 ♕xe2 24.♖d1 ♖d8
White is in zugzwang, and reduced solely
to pawn moves, which will eventually run
out – a harbinger of things to come! ]
21...♗h6 22.♗xe2
[ Bronstein points out the nice variation
22.♖e1 ♕g3 23.♖xe2?
and mates. ]
22...♕xc3 23.bxc3 ♖xe2
And now a delightful finish.
♖xd2! 25.♖xd2 ♖d8 26.♖hd1 c4!!
A gorgeous final position, reminiscent of
Nimzowitsch's famous 'immortal zugzwang
game' against S‫ن‬misch. White cannot move
a piece, without losing a rook on d2, and his
pawn moves will soon run out.
0-1
22
Bronstein,D
Ljubojevic,L
Petropolis Interzonal (Game 13)
[Steve Giddins]
This game is without doubt the most
tactically complex and fascinating in the
en t i r e b o o k . I t h a s b ee n th e s u b j ec t o f
detailed analyses by many top players,
including Bronstein himself, Vasiukov,
Timman, Speelman and Kasparov. In what
follows, I will try to bring together and
summarize the main points of these analyses,
with the aid (of course) of the computer, and
to help the reader get to grips with one of the
best games of the 1970s.
1.e4 ♘f6 2.e5 ♘d5 3.d4 d6 4.c4 ♘b6
5.f4 A good sign. White chooses the Four
Pawns Attack, undoubtedly the most critical
and dangerous (for both sides) response to
QUESTION: Is the Four Pawns really any
good? It does not seem to be very popular.
ANSWER: It is true that the line has not
been as popular in recent decades as some of
the quieter responses, but, in reality, I think
it is an extremely dangerous reply. I know of
one grandmaster, who a couple of years ago,
♕e1+! spent some time analysing Alekhine's
Defence, intending to play it as Black, but
24.♖d5 who eventually gave up on the idea,
precisely because he could not find anything
he regarded as satisfactory against the Four
Pawns.
dxe5 6.fxe5 c5? This move is the sharpest
of all Black's lines in the Four Pawns, and
has been a fascinating area for analytical
research over the years. However, it is
objectively bad.
QUESTION: Really? I thought it was
supposed still to be playable?
B03 ANSWER: Many books will tell you that.
However, my aforementioned grandmaster
friend revealed to me that according to his
1973 analysis (which I trust), 6...c5 actually loses
more or less by force.
QUESTION: Did he tell you how?
ANSWER: He did, as it happens, but that
particular trade secret will not be revealed in
this book.
[ 6...♘c6 is the sounder main line
alternative. ]
7.d5 e6
[ 7...g6 is objectively the only way to keep
Black on the board, I suspect, but I
cannot recommend it. ]
8.♘c3
[ Although 6...c5 is objectively not good, it
is a mass destruction weapon in online blitz.
I would not mind a glass of red wine for
every game that has gone 8.d6? ♕h4+
, winning for Black, for example! ]

Bronstein − Move by Move

 Bronstein − Move
48 by Move
8...exd5 9.cxd5 QUESTION: This position
looks like absolute suicide for Black! Surely
the central pawns are going to kill him?
ANSWER: Well, Alekhine's Defence is one
of the most archetypal of hypermodern
openings, and the whole idea is to encourage
White to set up a broad pawn centre, in the
hope that it will prove too unwieldy and will
serve as a target for Black's guerrilla raids
from the edge. The 6...c5 variation is an
extreme application of this approach and
certainly not for the faint-hearted.
c4
[ This was a relatively new idea at the time
of this game, replacing the discredited
older alternative 9...♕h4+ 10.g3 ♕d4
, after which 11.♗b5+ ♗d7 12.♕e2 ♘xd5
13.e6! fxe6 14.♕xe6+ ♘e7 15.♘f3 ♕f6
16.♕e2 was found to be good for White.
Indeed, one of the games referred to by
Kasparov, which reached this very position,
was Ljubojevic-Noses, Dresden 1969. As
we will see, the entire 6...c5 line became
a favourite battleground for the young
and highly talented Yugoslav GM. ]
QUESTION: So what is the point of 9...c4?
ANSWER: The move does several things. It
shuts in White's king's bishop, and it
prepares to bring its counterpart out, either
to c5, cutting through the heart of White's
position, or to b4, pinning the knight and
thus exerting pressure against the d5-pawn.
The whole of Black's play is about targeting
the two white pawns on d5 and e5 – if White
manages to maintain them and get castled,
then he will just have a crushing advantage,
so Black must fight tooth and nail against
this.
10.♘f3 ♗g4? Objectively, this has to be
condemned.
[ Modern theory prefers 10...♗b4 . ]
11.♕d4 Once again, the passage of time has
Bronstein − Move by Move
shown that this is not best.
[ Kasparov quotes 11.♗e2 as giving White
a clear advantage, but none of this was
known at the time of the present game. ]
11...♗xf3 12.gxf3 ♗b4 13.♗xc4 0-0
QUESTION: So what is going on here? The
position looks a real mess already.
ANSWER: Believe me, this position is a
picture of calm and order, compared with
what will follow. Both sides are playing
logically. Black has to get developed and
start creating threats against the white centre.
Already, in defending his two central passed
pawns, White has allowed the capture on f3,
splitting his pawns and exposing his king,
and it is still far from clear where the latter
will find a refuge – castling on either side
will leave it exposed, and in the centre it is
even more insecure. Black's immediate threat
is 14...Nc6 15 Qe4 Nxe5, demolishing the
white centre.
14.♖g1
[ White sets up threats along the g-file, and
with 14.♖g1 also prepares to meet ♘c6?
with 15.♕g4 , winning. ]
14...g6!
[ Ljubojevic's prior experience with this
line had also encompassed the white side
of this position. His opponent had played
14...♕c7? and been crushed after 15.e6 f6
( or 15...fxe6 16.♗h6 e5 17.♗xg7! exd4
18.♗xd4+ ♔f7 19.♖g7+ and wins –
Kasparov ) 16.♗h6 ♕xc4 17.♖xg7+ ♔h8
18.♖g8+! ♔xg8 19.♕g1+! and Black
resigned in Ljubojevic-Honfi, Cacak
1970. ]
15.♗g5! Even this was all theory at the time.
[ A correspondence game between two
British amateurs, Gibbs and Stewart, had
seen White try 15.♗h6? , but he ran into
♘c6 16.♕e4 ♘xe5! and was massacred.
Bronstein's move was a key improvement,
 Bronstein − Move
49 by Move

possibly prepared, although we cannot be accept the offered rook.

sure of this. ] 17...♗xg1!? QUESTION: Is this right?
15...♕c7 ANSWER: Probably. There are two
[ Previous annotators, starting with alternatives:
Vasiukov in the Russian tournament book, [ a) 17...♘8d7 was Vasiukov's
recommended 15...♕c8 as superior, so that recommendation. He then gives 18.d6 ♕c6
a later d5-d6 will not hit the queen. 19.0-0-0 ♗xg1 20.♖xg1 ♕c5 21.♖e1
However, Timman pointed out that this ♖ae8 as fine for Black, but Kasparov and
move has the crucial drawback of not the computer point out that, in this position,
attacking e5, with the result that White can simply 22.e6! fxe6 23.♖xe6 wins at once:
reply 16.♗b3 ( Kasparov gives the ♖c8 ( 23...♖xf4 24.♖xe8+ ♔g7 25.♖g8#
surprising 16.♖g4 , which he claims is also is mate ) 24.♖e5+ ♕c4 ( 24...♔h8
very strong) 16...♗c5 17.♕h4 . Now there 25.♖xc5 ♖xf4 26.♖xc8+ ♘xc8 27.♗xf4
is no 17...Qxe5+, as would have been l e a v e s W h i t e a p i e c e u p) 25.♕g3
available in the game. Timman then wins the black queen. ]
analysed only [ b) The other way of declining the rook is
A) 17...♕f5!? is much more tenacious, 17...♖e8 , once again going after White's e5-
but, the silicon monster finds the stunning pawn. Now Vasiukov once again had it
a n d v e r y s t r o n g m o v e 18.♖g2! wrong, claiming a refutation with 18.d6?
simply getting the rook off prise and ♗xd6 19.♘b5 , b u t m i s s i n g ♖xe5+!
preparing Ne4; White seems to be 20.♕xe5 ♗b4+ , winning for Black. ]
w i n n i n g h e r e : f o r e x a m p l e , ♘8d7 [ After 17...♖e8 Speelman instead gives
( 18...♕xe5+ 19.♖e2 ♕f5 20.♗c2 ♕c8 18.♗f6 , which led to a quick white win
21.0-0-0 gives a decisive attack) 19.f4! after ♘8d7 19.♘e4!
an d W h i t e h a s d e f e n d e d h i s c e n t r e , A) 19...♖xe5 is a better try, but the
repulsed all Black's threats, has an extra c o mp u t e r s h o w s t h a t 20.d6! ♗xd6
pawn, and plans simply 0-0-0, with a 21.♗xf7+ ♔f8 is winning; the star move
crushing attack; is 22.♕h6+! ( Kasparov condemns this
B) 17...♗xg1? 18.♕h6! , winning. ] line for White, but only considers
[ QUESTION: After Ljubojevic's move, 22.♗xe5 , which is unclear after ♗xe5
15...♕c7 , things look bleak for White, 23.♕h6+ ♗g7 24.♕xh7 ♘f6 25.♘xf6
don't they? After all, the bishop on c4 is ♕xf7 26.♕xg6 ♕xf6 ) 22...♔xf7
hanging and there is also a threat of 16... 23.♕xh7+ ♔e6 24.♗xe5 ♘xe5
Bc5, winning the rook at g1. What does 25.♖xg6+! ♘xg6 26.♕xg6+ ♔e7
White do? 27.♕f6+ ♔d7 28.♖d1 and White's attack
ANSWER: He sacrifices a whole rook, that is decisive: for example, ♘c4 29.♘xd6
is what. ] ♘xd6 30.♕f7+ ♔c8 31.♕f8+ ♕d8
16.♗b3! ♗c5 17.♕f4 32.♕xd6 , etc;
[ Note that 17.♕h4? , as in the previous B) 19...♘xe5 20.♖xg6+!! 1-0 in S.
note, loses here to ♕xe5+ . ] Marjanovic-B.Filipovic, Yugoslavia
After the text, we have a very critical 1974. ]
position. Black must decide whether to QUESTION: So the conclusion, after all

Bronstein − Move by Move

 Bronstein − Move
50 by Move
these complications, is that Black might as
well take the rook?
ANSWER: Yes.
18.d6
[ Ljubo's experiences with this line are
highly interesting. As we have seen, prior
to the present game, he had played the
white side a couple of times, and always
won. Then he switched to the black cause
for the present game, and lost. However,
undeterred, he carried on playing it as
Black, and racked up several victories. One
of these came later the same year, when the
Romanian GM Florin Gheorghiu took him
on, and attempted to improve on
Bronstein's play with the move 18.♔e2?
. However, this proved bad, and Gheorghiu
lost as follows: ♕c5! (as we will see below,
this is a key resource for Black in many
variations; the black queen defends the
bishop on g1 and, together, the two pieces
create threats against the white king)
19.♖xg1 ♕xg1 20.♗f6 (it looks as though
Black will be mated after Qh6, but the
white king is too exposed) ♕g2+ 21.♔e3
( 21.♔e1 ♘8d7 is no better: for instance,
22.♕h6 ♘xf6 23.exf6 ♖ae8+ 24.♘e4
♖xe4+ 25.fxe4 ♕xe4+ 26.♔d2 ♕d4+
and the f6-pawn drops) 21...♕xb2 22.♔d3
( or 22.♕h6 ♕c1+ ) 22...♘8d7 23.♘e4
♖ac8 24.♕h6 ♘xe5+ 25.♔e3 ♖c3+
0-1, F.Gheorghiu-L.Ljubojevic, Manila
1973. ]
Kasparov's 'My Great Predecessors' volume
reveals one other stunning thing about this
position. He claims that Ljubo recently
revealed to him that, after playing 18 d6,
Bronstein offered a draw, because he had
just five minutes (!) left on his clock. Ljubo,
who, as we have seen, was armed to the teeth
in this opening line, had used just ten
minutes on his own clock, so naturally he
Bronstein − Move by Move
declined. We can only be grateful that he did
so, as well as marvelling at Bronstein's
conduct of the rest of the game, with so little
time left.
18...♕c8? This proves to be a serious
mistake.
[ 18...♕c6? is no better, losing at once after
19.e6 fxe6 20.♗xe6+ ♔g7 21.♗h6+
and mate. ]
[ After 18...♕c6? 19.e6 , ♘8d7
is actually the best try, but also good for
White after 20.exf7+! ♔g7 21.0-0-0 .
QUESTION: How come a position like this
is good for White? He is still a rook down,
and what is he threatening? Say, I play
simply ♗c5 , getting the bishop back to
safety?
ANSWER: White still has a crushing attack,
and can take his time. Simply 22.♔b1!
gets the king off the c-file and prepares 23
Ne4 . The co mp uter sho ws th at Bl ack
c a n n o t d e f e n d : f o r e x a m p l e , ♖ac8
(pointless, but Black has no way to get any
pieces over to the kingside or to shut out
W h i t e ' s a t t a c k i n g u n i t s ) 23.♘e4
and there is no defence to 24 Bf6+.
QUESTION: It seems amazing that with so
many pieces on the board, and an extra
rook, Black cannot defend.
ANSWER: The problem is that his extra
rook is not really taking an active part in
the game, and the d6-pawn cuts his position
in half. Meanwhile, every white piece is
taking part in the attack, apart from his king.
If you have trouble believing that White
is winning here, it is worth spending some
time, pushing the pieces around and trying
to find a defence. You will learn a great
deal about the power of the initiative. ]
[ QUESTION: So if the other queen moves
lose, what should Black do?
ANSWER: The critical move is 18...♕c5!
 Bronstein − Move
51 by Move
, an idea we have already seen in the game
Gheorghiu-Ljubojevic, quoted above. Now
there are threats around the squares f2 and
maybe even e3. White continues 19.♘e4!
, when ♕d4! i s c r i t i c a l ( as shown by
Vasiukov, 19...♕e3+ 20.♕xe3 ♗xe3
21.♗xe3 ♘c6 22.f4 is good for White,
who has a pawn for the exchange and has
preserved his powerful central pawn
phalanx and dark-squared bishop ) .
QUESTION: And what is going on here?
ANSWER: The position is basically
unclear. Ljubo netted yet another point in
this variation, when he reached this very
position against the Israeli GM, Yehuda
Grünfeld, at the Riga Interzonal of 1979.
T h a t g a m e c o n t i n u e d 20.♖d1
( Timman suggested 20.♔f1 , but the
c o m p u t e r ' s ♘8d7 21.♖e1 ♘xe5
looks better for Black) 20...♕xb2 21.e6?!
( S p e e l m a n ' s 21.♖d2 forces a draw by
repetition, as the computer confirms: ♕a1+
22.♖d1 ♕b2 23.♖d2 ♕c1+ 24.♔e2
♘8d7 25.♗f6 ♘xf6 26.♘xf6+ ♔g7
27.♘h5+ gxh5 28.♕g5+ ♔h8 29.♕f6+
, etc ) 21...♘8d7! 22.e7! ♕xh2 23.exf8♕+
♖xf8 24.♕xh2 ♗xh2 25.♘f6+? ( 25.♔f2
is a better try, when Black is better but
not clearly winning) 25...♔g7 26.♘xd7
♘xd7 27.♗e7 ♖b8 28.♔f2 ♗e5 29.♖c1
♘c5 30.♖d1 ♗f6 31.♗xf6+ ♔xf6 32.♗c4
♖d8 33.♔g3 a6 34.♗f1 ♔e5 35.f4+ ♔e6
36.♗c4+ ♔f6 37.♔f3 b5 0-1, Y.Grünfeld-
L.Ljubojevic, Riga Interzonal 1979. ]
19.♔e2? Given the complexity of the game,
not to mention Bronstein's clock situation, it
is hardly surprising that the players should
exchange errors around here. Bronstein's
move looks logical, since it allows his rook
on a1 into the game, without pinning the
queen's knight (and thus threatens Ne4), but
it is actually wrong.
Bronstein − Move by Move
[ B r o n s t e i n p o i n t e d o u t t h a t 19.0-0-0!
i s c o r r e c t : f o r e x a m p l e , ♗c5?!
( Timman analyses the more tenacious
move 19...♕c5 when 20.e6 ♘8d7
21.exf7+ ♔g7 22.♔b1! is similar to the
line we analysed after 18...Qc6 – White
just unpins the knight and so threatens Ne4,
with a very strong attack; of course, it
was precisely to avoid such a tempo loss
with Kb1 that Bronstein chose instead to
put his king on e2, but this should have
cost him dearly) 20.e6 fxe6 21.♕e5! ♖e8
22.♗h6 ♕d7 23.♘e4 and the attack is too
strong. ]
19...♗c5? QUESTION: Another mistake!
ANSWER: Yes, and this time, it proves fatal.
Ljubo's error is perhaps less excusable, in
that he had plenty of time on his clock, and
apparently continued moving at near-blitz
speed. But even so, the complexity of the
game is just colossal, and mistakes are
inevitable in an over-the-board struggle,
even after long thought.
QUESTION: So what should he have
played?
[ ANSWER: The same key move that we
have already seen in several lines:
19...♕c5! . Now 20.e6? was analysed as
inadequate even in the pre-computer era by
Vainstein and Speelman, both of whom
pointed out the excellent defence ♘8d7!
( or 20...♕f2+ 21.♔d3 ♘8d7!
– Kasparov) 21.exf7+ ( 21.exd7 ♘xd7
leaves White with nothing; his deadly
central pawn duo has gone, and Black's
knight covers f6) 21...♔g7 22.♗h6+ ♔h8!
and the king escapes the checks, while
Black is still a rook up. ]
[ It looks as though 19...♕c5 walks into
20.♘e4 , but then there is a check on b5:
♕b5+ , after which it appears that Black is
winning. For example, 21.♔d2 ♘c4+
 Bronstein − Move
52 by Move
22.♔e1 and now Speelman's
A) Even 22...♗e3 , which Timman
analysed as losing to 23.♘f6+ ♔h8
24.♕h4 , is still winning for Black, after
♕b4+! 25.♔f1 ( or 25.♔e2 ♕d2+
26.♔f1 h5 and White is busted, since the
usual response 27.♘xh5 is just met by
♗xg5 ) 25...♘d2+ 26.♔e2 ♕xh4 .
QUESTION: So what are we concluding
here?
ANSWER: That 19...Qc5! is winning for
Black and refutes Bronstein's 19th
move.;
B) 22...♘e3! is one of several winning
moves. ]
20.♘e4 Now, on the other hand, White has a
winning attack.
♘8d7
[ 20...♘6d7 is slightly more stubborn, but
Timman gave a superb piece of analysis, to
s h o w h o w W h i t e s t i l l w i n s : 21.♖c1!
(the point of this move becomes clear an
i m p r e s s i v e t e n m o v e s l a t e r ) b6
( t h e c o m p u t e r ' s 21...♘c6 22.♘xc5
♘dxe5! keeps Black alive here, although
White still has a powerful attack for just
an exchange) 22.♗f6 ♘xf6 23.♘xf6+
♔g7 24.♕h4! h6 ( or 24...♖h8 25.♘h5+
♔g8 26.♗xf7+! ) 25.♘h5+ ♔h7 ( if
25...gxh5 26.♕f6+ ♔g8 27.♕g6+ ♔h8
28.♕xh6+ ♔g8 29.♗c2 f5 30.♗b3+ )
26.♕f6 ♖g8 27.♕xf7+ ♔h8 28.♕f6+
♔h7 29.♗xg8+ ♕xg8 30.♕e7+ ♔h8
31.♘f6 and wins. Note that, in this position,
if White's rook were not on c1, Black
would have a saving check on c4! ]
21.♖c1 QUESTION: What is the idea of this?
ANSWER: White sets up an exchange
sacrifice on c5, to lure the black knight away
from covering f6.
QUESTION: Once again, I am astonished
that White can afford such seemingly quiet
Bronstein − Move by Move
moves, when he is a rook down!
ANSWER: Yes, but as we pointed out before,
what is the extra black rook doing? The
answer is nothing – it sits on a8, contributing
precisely zero to the defence of the kingside.
In the area of the board that matters, namely
the black king's vicinity, White has every
one of his pieces, bar the king, taking part in
the attack. Even the rook on c1, as we have
said, is taking part, by threatening to lure
the d7-knight away from the critical zone.
♕c6 22.♖xc5! ♘xc5 White has now
invested two rooks for one minor piece, a
deficit of 'seven points' (reckoning a rook as
five points and a minor piece as three), but
again, what matters is the material that is
playing in the area of the board that matters,
i.e. the zone around the black king, There,
White is attacking with queen, rook, two
bishops and a knight, plus the e5-pawn
(which should certainly be counted, as it
controls f6, for example, as well as
threatening to advance to e6). Black is
defending with his king, rook and three
pawns, plus the knight at c5 – a hopeless
deficit.
QUESTION: But surely, this counting of
pieces is not serious, is it? It sounds like
something one uses with beginners!
ANSWER: It is deadly serious. Watch some
of the videos, in which Kasparov
demonstrates his games, and you will see
him employ exactly this approach. Yes, a
really experienced player tends to assess
such attacks more by 'feel' or instinct than by
actually counting, but that is only because he
subconsciously builds a similar counting
approach into his instinct in the first place.
For the less experienced, it is an excellent
way to assess whether an attack has a good
chance of succeeding.
23.♘f6+ ♔h8
 Bronstein − Move
53 by Move
[ 23...♔g7 24.♕h4 is no better. For
example, ♕b5+ 25.♔f2 h5 26.♘xh5+
and mate is forced: gxh5 27.♗f6+ , etc. ]
24.♕h4 ♕b5+ 25.♔e3!! Again, brilliant
tactical calculation by Bronstein, whose flag
must by now have been hanging.
[ T h e a p p a r e n t l y s a f e r 25.♔f2?
throws away the win after ♘d3+ 26.♔g1
♕c5+ 27.♔g2 ( 27.♔h1? actually loses
after h5! 28.♘xh5 ♕f2! 29.♘g3+ ♔g8
30.♗f6 ♕e1+ 31.♔g2 ♕d2+ 32.♔h1
♕c1+ 33.♔g2 ♘f4+ and Black will force
off the queens or mate White: for example,
34.♔f2 ♕d2+ 35.♔f1 ♘h5! 36.♘xh5
♖ac8 and wins ) 27...h5 28.♘xh5 ♕f2+
29.♕xf2 ♘xf2 30.♘f4 and White perhaps
has enough for the exchange, but no more. ]
25...h5
[ If 25...♕d3+ 26.♔f2 and now the d3-
square is blocked to the black knight. ]
26.♘xh5 ♕xb3+ QUESTION: Is giving up
the queen forced?
[ ANSWER: I am afraid so. After 26...♕d3+
27.♔f2 ♘e4+ ( 27...gxh5 28.♗f6+
mates ) 28.fxe4 ♕d4+ 29.♔f1 ♕d3+
30.♔e1 ♕b1+ 31.♗d1 the checks run
out. ]
27.axb3 ♘d5+ 28.♔d4 ♘e6+ 29.♔xd5
♘xg5 30.♘f6+ ♔g7 31.♕xg5 The smoke
finally clears, and White has a winning
material advantage. Now he just has to reach
the time control at move 40, which Bronstein
duly did.
♖fd8 32.e6 fxe6+ 33.♔xe6 ♖f8 34.d7 a5
35.♘g4 ♖a6+ 36.♔e5 ♖f5+ 37.♕xf5 gxf5
38.d8♕ fxg4 39.♕d7+ ♔h6 40.♕xb7 ♖g6
41.f4 An absolutely fabulous game, and one
o f th e mo s t i n t e r e s t in g e v er p la y ed b y
Bronstein. The complications were head-
spinning and even many of the later
published analyses, by such giants as
Timman and Speelman, contain mistakes,
Bronstein − Move by Move
sometimes ones which change the
assessment completely. As always, when one
subjects such pre-computer analysis to
silicon examination, errors turns up, but this
is no reflection on the players or analysts.
1-0
23 B20
Bronstein,D
Ratner,B
Dnepropetrovsk (Game 14) 1939
[Steve Giddins]
1.e4
[ What makes this game of special interest is
the parallel with one of Bronstein's most
famous games, or, mo re particularly,
opening ideas. Game 1 of his 1950
Candidates' play-off match against
Boleslavsky saw the players reach the
following position: 1.d4 ♘f6 2.c4 g6
3.♘c3 d5 4.cxd5 ♘xd5 5.e4 ♘xc3
6.bxc3 c5 7.♗c4 ♗g7 8.♘e2 0-0 9.0-0
cxd4 10.cxd4 ♘c6 11.♗e3 ♗g4 12.f3
♘a5 13.♗d3 ♗e6 . Bronstein now
produced the new idea 14.d5 , sacrificing
the exchange in exactly the same way as in
the present game. He went on to win, and
the exchange sacrifice itself has proved to
be one of his more enduring opening ideas,
which is still at the forefront of Grünfeld
theory to day. The present game shows that
the basic strategic idea was one which had
been in Bronstein's mind for many years
already, by the time he used it against
Boleslavsky. ]
1...c5 2.♘e2 QUESTION: A funny move!
ANSWER: It is an attempt to remain flexible.
White tries to retain the option of just
playing d4, and transposing back into main
lines, whilst also being able instead to
 Bronstein − Move
54 by Move
follow-up with g3 and play a kind of Closed
Sicilian structure. Paul Keres was one top
player who adopted it a few times.
d6 3.g3 ♗g4
[ 3...d5! , as pointed out by Fischer in 'My
60 Memorable Games', is quite effective
here, despite the tempo loss, since 4.exd5?!
♕xd5 is rather awkward for White. ]
4.♗g2 ♕c8 5.h3 ♗d7 QUESTION: I am
surprised by 5 h3. Now White is unable to
castle, as his h-pawn is attacked.
ANSWER: True, but Bronstein intends to
deal with this by a later g4. In such closed
positions, White often advances f4 and g4
anyway, so providing he waits until Black
has castled (thus reducing the force of the
reply ...h5), he should be fine; indeed, he
may well gain a tempo over the immediate ...
Bd7.
6.d3 g6 7.♗e3 ♗g7 8.c3 QUESTION: I
thought in the Closed Sicilian, the knight
usually came to this square?
ANSWER: In the main Closed Sicilian, that
is indeed the case, but the knight is not
always so well-placed there, since it later
gets hit by ...b5-b4, and Black can utilize the
d4-square for his knight. White often ends
up playing contortions such as Qd2, Nd1 and
then c3, to evict such a knight. Having the
option of putting a pawn on c3 is one of the
points of this 2 Ne2 set-up.
Positions with this arrangement were
explored in considerable detail by the
inventive Canadian IM, Lawrence Day, who
christened the set-up 'the Big Clamp' system,
and even published a pamphlet on it. British
GM Nigel Davies is another who has
extensive experience with such variations.
♘f6 9.♘d2
[ 9.♘a3 followed by Nc2 is another plan,
and the one generally preferred by
Lawrence Day. From c2, the knight
Bronstein − Move by Move
controls d4, and can later come to e3. ]
9...0-0 10.g4 As predicted, the moment
Black castles and takes his rook off the h-file,
Bronstein seizes the chance to advance g4,
setting his kingside pawns in motion and
preparing to castle.
♗c6!? QUESTION: What is this about?
ANSWER: I guess Black's logic was that,
now the bishop no longer has anything to do
on the c8-h3 diagonal, it redeploys to the
long diagonal, to oppose the white king's
bishop.
QUESTION: But I sense you disapprove?
ANSWER: Yes, I am not entirely convinced,
mainly because it leaves the queen's knight
without a decent square.
[ The move 10...♘c6 looks more natural,
although the black queen is not ideally
placed on c8 either, and will probably have
to lose a tempo to come to c7 at some
point. ]
11.0-0 QUESTION: So how do you assess
the position?
ANSWER: I prefer White's game, although it
is perhaps not a lot. He has a typical
development for these Big Clamp systems,
with his pawns on c3, d3 and e4 controlling
lots of central squares, and he has already
started advancing his kingside pawns, which
is his main plan here. Black, by contrast, is a
little slow. He has not yet completed his
development, and his pieces stand slightly
awkwardly, whilst his customary queenside
counterplay with ...b5-b4 has not yet got
going.
♘e8?! I do not especially like this either,
although it is true that it will soon be forced
after an eventual g5. I should prefer to get on
with the queenside development and start
some counterplay on that flank.
12.♘g3 ♘d7 13.f4
[ QUES TION: W a s n' t 13.d4 possible
 Bronstein − Move
55 by Move
here?
ANSWER: Yes, and it is a perfectly decent
move, which shows how Black has failed
to develop very purposefully and control
the d4-square. But it would also allow the
c-file to be opened, and Bronstein does
not wish to give his opponent that chance
of counterplay. ]
13...♘c7 14.♘f3 b5 15.d4 QUESTION: So
he does it anyway!
ANSWER: Yes, but now Black has an extra
minor piece on the c-file, so it will take some
time for him to threaten any sort of heavy
piece penetration down the open c-file,
should he exchange on d4.
QUESTION: White looks to have a great
position!
ANSWER: It is probably not as great as it
looks, but he is certainly better. I am
surprised that Fritz only assesses the position
as equal. I think that says rather more about
the computer's judgement in such amorphous
structures than it does about the position
itself.
♖b8 16.b4!? QUESTION: That is a
surprise! I thought one was not supposed to
move pawns on the side where the opponent
is attacking?
ANSWER: That is a frequently-cited rule of
thumb, but, like every such rule in chess, it
has its exceptions. Bronstein's idea is to kill
off Black's usual plan of ...b4, when his rook
on b8 would be well-placed, and there would
be the idea o f o penin g the b -file by an
exchange on c3, and then trying to penetrate.
QUESTION: But now Black can open the c-
file instead.
ANSWER: Yes, but as we have already noted,
the logjam of minor pieces on the c-file
makes it difficult for him to exploit that file.
QUESTION: So you like 16 b4 then?
ANSWER: Yes, I do. The main drawback is
Bronstein − Move by Move
that it leaves a hole on c4, which Black's
knight can seek to occupy by ...Nb6-c4, but
even if he does so, it is not clear how he
follows up. The knight would look nice on c4,
but it also blocks the c-file for Black's heavy
pieces, and on its own, the knight does not
actually achieve a great deal. The other point
a bout 16 b4 i s t ha t i t pre pa re s Wh i t e ' s
striking 18th move, which Bronstein had
probably already seen.
cxb4 17.cxb4 ♘b6 A logical enough
response, but now the 15-year old Bronstein
uncorks a cracking idea.
18.d5!? QUESTION: Giving up the
exchange?
ANSWER: Yes. This is a purely positional
sacrifice, with no immediate mate or
recovery of material. Bronstein simply
believes that the black fianchettoed bishop is
more important than the white rook on a1.
♗xa1 19.♕xa1 ♗d7 QUESTION: So what
does the computer make of this position?
ANSWER: As you might expect, the silicon
miser is not convinced by Bronstein's
exchange sacrifice, but I really do not think
one can take too much notice of that. The
same is true of the Grünfeld variation above,
which, in the absence of its opening book,
the computer assesses as better for Black.
The long-term nature of the sacrifice, and the
absence of direct threats, means that the
computer will always struggle to appreciate
such sacrifices.
[ When it comes to concrete variations,
though, the silicon beast is much, much
stronger, and here it suggests as Black's
b e s t m o v e t h e k n i g h t j u m p 19...♘c4
. White then has a choice between the
s i m p l e r e t r e a t 20.♗c1 ( and grabbing a
pawn with 20.♗xa7 ♖a8 21.♗d4 ♗d7 )
20...♗d7 21.f5 with play similar to the
game. In either case, the computer claims
 Bronstein − Move
56 by Move
a clear advantage for Black, but I am not
convinced this assessment can be entirely
trusted, for the reasons outlined above. I
think White has compensation here. ]
20.f5 Now even the computer thinks White
has sufficient compensation to balance out
the material, which I think bears out my
previous comment.
QUESTION: How do you mean?
ANSWER: Well, the only difference in the
position from the variation after 19...Nc4 20
Bc1 Bd7 21 f5 is that the white bishop is on
c1 instead of e3, and the black knight on c4
instead of b6. I agree that these changes are
slightly in Black's favour, since his knight is
more actively placed, but I really don't see
that the difference is that great.
f6 21 Bh6 was threatened, which would
regain the exchange and force this
weakening move anyway, so Black had little
choice.
21.♗h6 ♖f7 This results in White regaining
the exchange, thanks to a tactic, but Black
did not have a lot of choice.
[ If 21...♖e8! he leaves the f6-square too
vulnerable, and after 22.g5! the defensive
barriers around the black king prove to
have all the strength and resistant
qualities of damp polystyrene.
The other important point to note about
this position is the way the central pawn
configuration, especially the pawn on e7,
cuts Black's position in half, and prevents
his pieces on the queenside from getting
across to help in the defence of their king.
Black's queen, rook, bishop and two
knights are all pretty well sealed off from
the kingside, whilst White is attacking on
that wing with almost his entire army.
QUESTION: We have seen this scenario
before in this book!
ANSWER: We certainly have, and it is so
Bronstein − Move by Move
often the key to understanding how an
attack can s ucceed . I t is n ot the to tal
material balance on the board as a whole
that matters, but the balance in the key
sector, where the action is taking place. ]
22.fxg6 hxg6 23.♘g5! ♘e8?
[ The knight obviously cannot be taken, in
view of forced mate: 23...fxg5? 24.♖xf7
♔xf7 25.♕g7+ ♔e8 26.♕f8# . ]
[ However, although Black must lose back
the exchange, he does have a major
decision to make, in terms of which white
minor piece he should get for the rook.
Ratner's choice was the knight on g5, but
the computer heartily condemns this and
advocates 23...♖g7 instead, choosing to
eliminate the white queen's bishop.
QUESTION: What do you think?
ANSWER: It is not an easy decision over
the board, but the variations seem to
s u p p o r t t h e c o mp u te r 's c h o i c e . A f t e r
24.♗xg7 (not obligatory, but I am not sure
White can gain anything much from
d e f e r r i n g t h e m o v e ) ♔xg7 25.♘e2
White retains the initiative, but Black is
probably OK. It seems this was his only
hope, as the game continuation sees his
position collapse quickly. ]
24.♘xf7 ♔xf7 25.e5! This is the real killer
follow-up, which Black probably missed or
underestimated. The pressure against f6 and
the other dark squares around the black king
is too great to resist.
dxe5 26.♘e4?!
[ A creative follow-up, but the computer's
straightforward 26.♕xe5 was probably
stronger. Now Black gets another chance to
hold his position together. ]
26...♘c4 27.♘g5+ ♔g8 28.♕b1
QUESTION: What is going on now?
ANSWER: This last move was the point of
Bronstein's idea. He attacks g6, which
 Bronstein − Move
57 by Move
cannot be defended other than by the advance
of the f-pawn, which obviously weakens
Black's position enormously. However, it
seems he can still defend.
f5 29.gxf5 gxf5
[ 29...♗xf5! was better, activating the black
pieces. Then after 30.♕e1 ♘f6 31.♕g3
♘h5 Black can hope to repulse the
threats. ]
30.♕d3 ♕d8? This proves fatal, by allowing
White's next move.
[ Black had to blockade the d5-pawn with
30...♘ed6 , when the complicated variation
31.♘e6 ♗xe6 32.♕g3+ ♔f7 33.♕g7+
♔e8 34.dxe6 ♕xe6 35.♕f8+ ♔d7
36.♕xb8 ♕xh6 ends in perpetual check
after 37.♖xf5! ♘xf5 38.♕xb5+ ♔d8
39.♕b8+ ♔d7 40.♕b5+ etc, since if ♔e6
41.♕xc4+ ♔f6 42.♕a6+ e6 43.♕xa7
♕c1+ 44.♔h2 ♕f4+ it is Black who gives
perpetual. ]
31.d6! Clearing the path for the light-squared
bishop to enter the game, with decisive effect.
It is a nice feature of the game that the
attack initially concentrated on the dark
squares around the black king, but then
moves such as 28 Qb1 and now 31 d6 switch
it to the light squares.
♘cxd6
[ Q U E S T I O N : W h a t h a p p e n s i f 31...e6
to keep the light-squared bishop out?
ANSWER: Then the computer shows a
b r i l l i a n t w i n b y 32.♖xf5! , crashing
through on those same light squares. For
example, exf5 33.♗d5+ ♔h8 34.♘f7+
♔h7 35.♘xd8 ♖xd8 and now simply
36.♗g5 ♖c8 37.♕g3 finishes Black off in
short order (it is mate in four, in fact). ]
32.♗d5+ ♔h8
[ 32...e6 33.♖xf5! is another brilliant
finish: ♘xf5 34.♕xf5 ♕e7 35.♕g6+
♔h8 36.♘f7+ , etc. ]
Bronstein − Move by Move
33.♕g3! Calmly transferring the queen to h4,
where the threats will be decisive.
e6 34.♘xe6 ♕e7
[ 34...♗xe6 35.♗xe6 leaves Black unable
to cover g8, without unprotecting g7. ]
35.♘f8 Black resigned, as again he cannot
protect the g8-square and avoid 36 Ng6+ as
well. The finish is a lovely early example of
Bronstein's attacking prowess, but what
lends the game special interest is the striking
exchange sacrifice on move 18, with its eerie
presaging of Bronstein's later famous
Grünfeld innovation.
1-0
24 B83
Bronstein,D
Levenfish,G
USSR Championship, Moscow (Game 15)
[Steve Giddins]
1.e4 c5 2.♘f3 ♘c6 3.d4 cxd4 4.♘xd4
♘f6 5.♘c3 d6 6.♗e2 At the time of this
game, this was the main move here, but in
view of the strength of Black's reply, it fell
out of fashion.
[ For a long time, White used to play the
Sozin line 6.♗c4 , ]
[ but that in turn has now given way
overwhelmingly to Rauzer's 6.♗g5 . ]
6...e5 The Boleslavsky System, the
granddaddy of all these ...e5 lines in the
Sicilian.
QUESTION: It was quite a revelation in its
day, wasn't it?
ANSWER: Yes, it revolutionized our
understanding of such structures.
Boleslavsky (whose daughter eventually
married Bronstein) showed that Black's
dynamic play compensates for the apparent
weakness of the d5- and d6-squares.
 Bronstein − Move
58 by Move
7.♘b3 ♗e7 8.0-0 ♗e6 9.f4 exf4 10.♗xf4
0-0 QUESTION: How would you assess the
position?
ANSWER: Chances are equal. White has
more space, because of his central pawn on
the fourth rank, and he still potentially has
an outpost on d5, and a weakness to attack
on d6. But he also has a weak pawn on e4,
and Black has the half-open c-file. Black's
king is also a little more secure, White's
h a v i n g b e e n e x p o s e d s o me w h a t b y t h e
advance f4.
11.♔h1 d5?! QUESTION: Why is this
wrong? Surely, ...d5 is the advance Black
usually wants to play in the Sicilian, and it
gets rid of his weak pawn.
ANSWER: All of that is true, and at first
sight, one would expect Black to be doing
fine if he can get in ...d5. But Boleslavsky
himself, who annotates the game in the
Russian tournament book, criticizes the move
as "not in the spirit of the position" and
allowing White to obtain a positional
advantage.
[ I n s t e a d , h e r e c o m m e n d s 11...♕b6
, in t en d in g . .. N e5 , a n d th e n i f 12.♘d5
♗xd5 13.exd5 ♘e5 14.c4 ♘e4
, as had all been played two years earlier in
a game A.Sokolsky-D.Bronstein, Leningrad
1947. ]
12.e5 ♘e4
[ Boleslavsky mentions 12...♘d7 13.♘xd5
♘dxe5 14.c4 , and assesses the position as
highly favourable for White. That seems
an exaggeration, but White has an edge,
thanks to his queenside majority. ]
13.♗d3! f5 QUESTION: I don't understand
this. Why not double White's c-pawns?
[ A N S W E R : A f t e r 13...♘xc3 14.bxc3
the pawns are less weak than they look,
and meanwhile, White has a nice outpost
on d4 for his knight. Even so, chances are
Bronstein − Move by Move
probably only equal, although Boleslavsky
gives the impression he thought White
was better here. ]
14.exf6 ♘xf6!?
[ Boleslavsky rejects 14...♗xf6 because of
15.♘xe4 dxe4 16.♘c5 "winning a pawn",
but the computer continues the variation
with exd3 17.♘xe6 dxc2! 18.♕xc2
( g r a b b i n g t h e e x c h a n g e b y 18.♕xd8
♖fxd8 19.♘xd8 ♖xd8 is actually bad for
White, because the c2-pawn is too strong:
for example, 20.♖ac1 ♘b4 21.a3 ♘d3
22.♖xc2 ♘xf4 wins material, due to the
back-rank weakness) 18...♘d4 19.♘xd4
♕xd4 , with dead equality. ]
15.♕e1 ♕d7 16.♘a4 ♗f7? Now Black gets
into trouble.
[ The threat of 17 Nac5 looks unpleasant,
but typical Sicilian dynamics come to his
aid: 16...♘e4! with the point that 17.♗xe4
dxe4 18.♕xe4 ♗d5 19.♕e2 ♖ae8
gives Black adequate compensation for the
pawn, in the shape of his two strong
bishops and the exposed white queen. ]
17.♘ac5 ♕c8 18.♗g3
[ 18.♗g5! was stronger, as pointed out by
Boleslavsky, with ideas of Bf5 and also
Qh4, attacking the kingside. ]
18...♗xc5!?
[ Surrendering the bishop-pair is a long-
term concession, but I am not convinced
that the computer's malodorous suggestion
of 18...♗g6 is really any better. After
19.♗xg6 hxg6 20.♖d1 the weak black
pawns should guarantee White a lasting
edge. ]
19.♘xc5 ♖e8 20.♕f2 QUESTION: Things
seem to have gone a little wrong for Black?
ANSWER: Yes, now White has a nice, and
long-lasting positional advantage. He has
two bishops, and excellent control of the
weakened dark squares in Black's position.
 Bronstein − Move
59 by Move
♘e5 21.♗f5 ♕c6 22.♖ae1 ♘eg4
23.♕d4!? Probably not the most accurate,
although White is still better.
[ Boleslavsky says 23.♖xe8+ ♕xe8 24.♕f4
♘h6 i s w e a k e r , b u t a f t e r 25.♗d3
White keeps a solid advantage,
( whereas Boleslavsky only considers the
weaker 25.♘xb7 . )]
23...b6?! QUESTION: You don't like this?
ANSWER: I think this time Boleslavsky was
right on the mark, in saying that the text
leads to the exchange of queens and a
difficult endgame for Black.
[ His recommendation is also the computer's
first choice: 23...♘h6 24.♗d3 b6 25.♘b3
♘e4 "with some counterplay for Black".
W h i t e r e m a i n s b e t t e r a f t e r 26.♗f4
, but a position with queens on offers
Black more chances of creating counterplay
than a lifeless technical ending. ]
24.♘d7 ♕c4
[ 24...♘xd7 25.♕xg4 ♘f6 26.♕h3
is also unpleasant, but would have the
same merit as the line in the previous note,
of keeping queens on. ]
25.♘xf6+ ♘xf6 26.♗e5 ♕xd4 27.♗xd4
QUESTION: So White has the two bishops
in the endgame.
ANSWER: Yes, and this position is solid
masochism for Black. As well as the two
bishops, White also has one fewer pawn
island. The computer may assess White's
advantage as no greater than in the previous
couple of notes, where the queens stayed on
the board, but in practice, I think the game
position is much harder and more unpleasant
to defend as Black.
QUESTION: So why did Levenfish play this
way?
ANSWER: That is a good question. He was
himself a great endgame specialist and a
decade or so later, he wrote a classic book on
Bronstein − Move by Move
rook endings (it appeared under the joint
names of himself and Smyslov, but the
latter's name was only added in order to get
the book in print; Levenfish had always been
viewed with suspicion by the Soviet
authorities, as he had a pre-Revolutionary,
bourgeois background).
QUESTION: So, all the more reason why he
should have known better!
ANSWER: Yes, but paradoxically, this may
have been precisely why he entered this
ending. I suspect Levenfish thought that a
young master such as Bronstein, who was
known for his fiery imagination and tactical
play, would find it much harder to outplay
such an experienced endgame expert as
himself in a position like this, whatever the
objective merits of White's position. But if
so, he proved very wrong, as Bronstein gives
a textbook demonstration of how to handle
such positions.
♘e4 28.♔g1 ♗g6 29.♗d7 QUESTION: He
turns down the chance to double Black's
pawns?
ANSWER: Yes, there would be no value in
doing that, if it costs White his bishop-pair.
QUESTION: But he would still have bishop
versus knight.
ANSWER: Yes, but that is usually
significantly less of an advantage than two
bishops against bishop and knight.
QUESTION: Why?
ANSWER: Because the bishop itself has one
great disadvantage against a knight, even in
an open position – it can only control
squares of one colour. So, if White were to
exchange light-squared bishops here, his
remaining bishop on d4, splendid piece
though it is, can only control half the squares
on the board, whereas the black knight can
go to all 64. This 'colour-blindness' of a
bishop is a major issue, which is often
 Bronstein − Move
60 by Move
overlooked. However, two bishops between
them control every square on the board, of
course. So the net result is that two bishops
together are often worth more than the sum
of the parts, and exchanging off one of them
is a standard defensive technique.
♖e7 30.♗a4 The bishop transfers round to
the excellent square b3, from where it takes
aim at the weak d5-pawn.
♖f8 31.♖xf8+ ♔xf8 32.♗c6 ♘f6
[ Practically forced to save the d5-pawn,
since after 32...♗f7 33.c4 , Black loses
material. ]
33.♖xe7 ♔xe7 QUESTION: How has the
exchange of rooks affected things?
ANSWER: It has not really changed much;
certainly it hasn't diminished White's
ad vantag e. He no w has th e clas sic two
bishops versus bishop and knight ending,
with an extra target on d5. Theoretically, his
advantage may not be that great, but Black
rarely manages to hold such positions, when
a leading GM is playing White.
34.c3 Threatening to win a pawn by taking
on f6 and d5.
♔d6 35.♗b7 ♗b1 36.a3 ♗d3 37.♔f2 ♗e4
QUESTION: Black seems just to be waiting.
ANSWER: He has nothing else to do. In
most cases, thrashing about in such a
position will only make things worse, if
there is no real enemy weakness to go after.
Black just has to sit tight and hope he can
hold his lines.
QUESTION: And what will White do? It is
not obvious how he can make progress – d5
is securely defended and there are no other
black weaknesses.
ANSWER: The standard plan in such
positions is to bring the king as near the
centre as possible, and then start advancing
the wing pawns, to gain space and drive the
enemy knight out of any decent central
Bronstein − Move by Move
outposts.
38.h3 And that is precisely what Bronstein
now proceeds to do. He plans to push his g-
pawn.
♘e8 39.g3 QUESTION: That doesn't look
very dangerous! I thought you meant he
would push it to g4.
ANSWER: I did, and he will.
[ Q U E S T I O N : S o w h y n o t p l a y 39.g4
at once?
ANS WER: It is typ ical of the 'd o no t
hurry' principle, which is so important to
good endgame technique. There is no rush,
so White first nudges his pawn forward,
getting it off prise from the bishop, and
then waits to decide when to advance it
further. ]
39...♗f5
[ QUESTION: But what if Black plays
39...h5 now?
ANSWER: That would create a serious
weakness. White could fix the pawns with
40.h4 and then try to bring his king into g5.
There is one other factor behind
Bronstein's slow play hereabouts. In those
days, games were adjourned after 40 moves,
and only resumed several days later. It
was thus standard technique not to take
crucial or committal decisions in the last
few moves of the first session, but to
postpone them until after adjournment,
when one could analyse them at leisure. ]
40.g4 ♗e6 41.♗e3 ♔c7 42.♗a6 ♘d6
43.♗d3! QUESTION: Why is this good?
ANSWER: It is a typical weakness-
provoking move. The attack on h7 forces
Black to do something. If he advances the h-
pawn, he creates a weakness on g6, and if he
advances the g-pawn, he creates one on h6/f6.
Provoking weaknesses like this is an
essential part of endgame technique.
g6
 Bronstein − Move
61 by Move
[ 43...♘e4+ avoids either weakening pawn
move for the moment, but the knight is
unstable on e4 after 44.♔f3 . The threat of
c4 hangs in the air. ]
44.♗c2 ♘c4 45.♗c1 QUESTION: It looks
as though Black is fighting back?
ANSWER: No, this is just an illusion. Yes,
for the moment, the white bishops appear
passive, but the knight can be driven away
again with b3, whenever White wants, and
meanwhile, nothing essential has changed.
♔d7 46.♔g3 ♔e7 47.♔f4 See the note to
White's 43rd move. Having provoked the
move ...g6, weakening the dark squares on
the kingside, White attempts to infiltrate on
those weak squares with his king. Black
must prevent Kg5.
♔f6 48.h4 Stopping Black playing ...g5+
himself, and asking him what he is going to
do next.
QUESTION: It all seems so slow, like White
is fighting for trifles!
ANSWER: That is the nature of endgame
technique. Slow, systematic probing, and the
accumulation of tiny advantages is the order
of the day. Eventually, the straws will mount
up, until one of them breaks the camel's back.
♗d7 49.b3 ♘e5!?
[ 49...♘d6 would avoid the tactic that
follows, but it does not essentially change
anything. White can just continue with
50.♗e3 , preparing a check on d4, and if
♘b5 then 51.a4 ♘xc3? 52.♗d4+ . ]
50.c4! QUESTION: A tactic at last!
ANSWER: Indeed. Another aspect of good
endgame technique is not to forget about
tactics. The play may be slow and full of
p r o t r a c t e d ma n o e u v r i n g , b u t t h e g r e a t
endgame players never forget about tactical
opportunities, and make full use of them
when they appear. Here, the threat of 51 Bb2
forces a concession from Black.
Bronstein − Move by Move
♔e6 QUESTION: Can't he take on g4?
[ ANSWER: He can, but after 50...♘xg4
51.cxd5 White has a passed pawn, which
will be a serious trump card. ]
51.♗b2 ♘f7 52.cxd5+ ♔xd5 QUESTION:
I don't understand what White has achieved
over the last few moves. Hasn't he just
exchanged off Black's weak d5-pawn for
him?
ANSWER: Yes, but he has also opened more
lines and cleared more space for his pieces.
The d5-pawn was never in real danger of
dropping off, but it was depriving the white
pieces of some squares and lines that would
have been useful. Now, for example, the a2-
g8 diagonal is open, so Black must worry
about the white bishop getting round the
back to g8.
53.♗e4+ ♔d6 54.♗g7 QUESTION: Now
what is he doing?
ANSWER: The bishop just controls some
squares, preventing a possible ...Nh6,
attacking g4, and also setting up a potential
Bf8+, driving the black king back.
QUESTION: You say "a possible" this and
"a potential" that. You don't seem sure if
these are threats or not!
ANSWER: No, I am not. That, again, is how
endgames work. Bf8+ may never happen, but
it might. Either way, it is yet another thing
for Black to worry about. The possibility of
Bf8+ may persuade him to play in a different
way from otherwise, and White may then be
able to exploit that. It is all part of the
process of just grabbing every little
advantage that one can – to repeat the earlier
analogy, one never knows which straw will
be the one that breaks the camel's back, so
one just goes on gathering as many as
possible, and waits to see.
♗e6 55.b4 ♔e7 56.♗d3 ♘d6 57.h5
Now that the knight has left f7, there is no
 Bronstein − Move
62 by Move
longer the move ...g5+ for Black, so White
seizes the opportunity to attack his kingside
pawns.
♗f7
[ 57...gxh5 58.gxh5 leaves the h7-pawn
very hard to protect, and Black would be
f o r c e d i n t o t h e h o r r i b l y p a s s i v e ♗g8
. Rather than that, Black prefers to defend
a pawn on g6, but either way, his position
is becoming progressively more
uncomfortable. ]
58.♔g5 Now he has to take on h5 anyway.
gxh5 59.gxh5 ♗g8 60.h6 Fixing the pawn
on h7 and thus ensuring that the black
bishop will be tied down on g8 for the
foreseeable future.
QUESTION: But I still don't see how White
is going to win.
ANSWER: Nor do I, and, frankly, I don't
suppose Bronstein did yet, either. But he is
making steady progress and has seized some
further plusses, in the shape of creating a
new weakness on h7 and tying down the
enemy bishop. Now he will consider how to
progress further.
b5 Black tires of waiting passively and tries
to set up some counterplay with ...Nc4.
QUESTION: What happens if he waits?
ANSWER: Well, as I say, there are various
possibilities, but the one that probably
provoked the text is the idea of a4, followed
by Bd4 and a5. That would create a parallel
situation on the queenside to what we just
saw on the kingside, but this time, Black
would not have the pieces with which to
defend the queenside.
61.♗d4 ♘c4 QUESTION: Doesn't that lose
a7?
[ ANSWER: The trouble is, Black cannot
afford the time for 61...a6 because of
62.♗c5 and there is no defence to the threat
of taking on d6 and entering on f6-g7 with
Bronstein − Move by Move
the white king: ♔e6 63.♗xd6 ♔xd6
64.♔f6 , etc. Instead, Black sets his hopes
on counterattacking against a3. ]
62.♗c5+! QUESTION: Why is this good?
ANSWER: It worsens the position of the
black king.
QUESTION: But how?
ANSWER: Well, where are you going to go?
[ QUESTION: What about meeting
62.♗c5+ with ♔e8 ?
A N S W E R : T h e n I h a v e 63.♔f6
, and Kg7 will decide. ]
[ QUESTION: Well, OK then, I will try
62.♗c5+ ♔f7 .
ANSWER: That has the same problem as
the move in the game – it blocks the action
of Black's bishop, and allows 63.a4 . ]
62...♔e6 63.a4! Exploiting the fact that the
black king has blocked its bishop from
defending the knight.
a6 64.a5! Now the pawn on a6 is another
weakness. White threatens Be4 and Bb7,
hence Black's next move.
♔e5 65.♗f8! QUESTION: Now what?
ANSWER: White prepares a check on g7.
Then the black king will not be able to cover
both f6 (to stop Kf6-g7) and e4 (to stop Be4-
b7). The end is near.
♘e3 66.♗g7+ ♔d6
[ T h i s a l l o w s B e 4 , b u t a f t e r 66...♔d5
67.♗c3 White threatens Kf6-g7. ]
67.♗e4 ♘d5 The only way to keep the
enemy bishop from b7. Black is clinging on
desperately by his fingertips, but White will
inevitably find a way to stamp on his fingers.
68.♗f8+ ♔c7
[ Q U E S T I O N : W h a t a b o u t 68...♔e6
, keeping the white king out?
ANSWER: Then simply 69.♗f5+ ♔e5
70.♗c8 ♘c7 71.♗g7+ ♔e4 72.♗a1
and again, Kf6-g7 will follow. Black is
unable to defend the two far-flung
 Bronstein − Move
63 by Move
weaknesses on a6 and h7. ]
69.♔f5 ♘c3 70.♗c2 ♘d5 71.♔e5
White intends Be4, forcing the knight away,
and then Kf6.
♘e3 72.♗e4 ♘g4+ 73.♔f5 ♘f2 74.♗c2
Now there is finally no defence to 75 Kf6.
♗c4 75.♔f6 ♗d3 76.♗xd3 A masterly
technical display by Bronstein. A careful
study of this ending will teach you more than
any half-dozen opening books put together.
It also casts an interesting light on the oft-
heard claim th at Bro nstein was a weak
endgame player. This particular canard
gained widespread acceptance after his
match with Botvinnik, where he lost several
half-points through mistakes in endgames,
but it was a great exaggeration of the truth.
As this game shows, Bronstein could play
the endgame as well as anyone, when he was
in form. I can think of plenty of modern-day
GMs, who wish they wer e as 'weak' as
Bronstein in endgames.
1-0
25 B51
Sax,G
Bronstein,D
Teesside (Game 16) 1975
[Steve Giddins]
1.e4 c5 2.♘f3 d6 3.♗b5+ This quieter
alternative to the main lines has always
retained a measure of popularity and has
been used a few times by our current theory-
averse world champion.
♘d7
[ 3...♗d7 ]
[ and 3...♘c6 are both good alternatives,
but the text is perhaps a little more
combative, certainly than the former.
Black aims to avoid too many exchanges
Bronstein − Move by Move
and ensures that, after a subsequent ...a6,
White will either have to retreat the bishop,
with potential loss of tempo, or surrender
the two bishops. In both cases, Black hopes
the game will have more tension than in the
3...Bd7 variation. ]
4.d4
[ 4.c3 is a quieter alternative, aiming for a
type of Spanish build-up, but the text is
also a main line. White's idea is to exploit
the absence of the enemy knight from c6,
by recapturing on d4 with the queen. ]
4...♘gf6 5.♘c3 cxd4 6.♕xd4 e5
QUESTION: That is a surprise. What about
the backward pawn on d6?
ANSWER: 6...e5 is typical of a move which
would probably have horrified the pre-WW1
generation of Tarrasch and Teichmann, but
which is now quite standard. In the Open
Sicilian (which pawn structure we have now
reached), Black has an extra centre pawn,
and one of the most important developments
in chess in the 20th century, especially after
WW2, was the realization that Black can use
this extra centre pawn to stake a claim for
central space and control.
On e5, the pawn controls d4 and f4.
Admittedly, the pawn on d6 is backward and
the d5-square a 'hole', but the efforts of
pioneering researchers, especially
Boleslavsky, showed that Black's dynamism
can prevent White from settling down to a
cosy exploitation of these static weaknesses.
This understanding is at the heart of some of
the most popular Sicilian variations,
especially the Najdorf and Sveshnikov
systems. We saw an example in the previous
game, of course, where Black was fine out of
the opening.
7.♕d3 h6! QUESTION: Why is this so
good?
ANSWER: Black prevents the move Bg5,
 Bronstein − Move
64 by Move
which would exchange off the knight on f6,
which defends d5.
QUESTION: But the other black knight can
just recapture.
ANSWER: Yes, but in general, in these ...e5
structures, Black is relying on dynamism,
and so he prefers to avoid too many
exchanges. The latter would tend to suck the
life out of the black position, and could
eventually see his static weaknesses assume
greater importance. The two black knights
can both control d5, whereas the white
queen's bishop cannot, so it makes sense to
prevent the exchange of bishop for knight.
QUESTION: OK, but even so, 7...h6 does
not develop a piece. Can't White exploit this?
ANSWER: Not really, because the position
is actually not that open. White has no
immediate pawn levers and cannot prevent
Black completing his kingside development
in peace.
8.♗e3 ♗e7 9.a4 0-0 10.0-0 ♘b8!
QUESTION: What on earth is he doing?
Why undevelop a developed piece?
ANSWER: Bronstein was very pleased with
this move and it became quite trendy in such
positions. His point is that the knight on d7
is not ideally placed, as it blocks the queen's
bishop, which would like to come to e6,
guarding the vital d5-square. In addition, the
queen's knight itself would be more active on
c6.
QUESTION: But what about Black's
development? He is already a long way
behind, and this makes things even worse.
ANSWER: Yes, but once again, the issue is
levers. An inferiority in development can be
serious, even fatal, if the opponent can
exploit his superior mobilization to launch a
quick attack. But here, we have quite a static
structure. White does not have any way to
lever open lines, so he is not able to exploit
Bronstein − Move by Move
his temporary development advantage.
11.♗c4 ♘c6 12.h3 ♗e6 13.♗xe6 fxe6
QUESTION: So what is going on here? As
well as a backward d6-pawn, Black has
doubled e-pawns. Surely this must be bad for
him?
ANSWER: Actually, not! Once again, it is
our old friend, the 'unexploitable weakness'.
The pawns on e6 and d6 look as though they
should be weak, but White is not able to
attack them successfully. How can he attack
e6, for example? Only by Qc4, but that is
easily met by ...Qd7, and then the white
queen will soon be driven from its exposed
position on c4. Likewise, the d6-pawn can
be attacked by Rad1 and Nb5, but Black can
defend it by ...Ne8, if necessary, and then ...
a6 will drive off the enemy knight, after
which White is not able to get any more
pressure against d6.
QUESTION: It is all very confusing! How
are we supposed to know when a 'weak' pawn
is weak and when it isn't?
ANSWER: Well, I never said chess was an
easy game, but the issue is one of specifics.
A formal weakness, such as a backward or
doubled pawn, is only a weakness if it can be
effectively attacked. As it is, the black
central pawn formation here also performs a
very useful function. If we look at the four
main central squares, e4, d4, d5 and e5, we
see that Black controls or occupies three of
them, whilst his knight on f6 attacks the
other one (e4) and forces White to defend it.
Furthermore, Black is constantly threatening
to advance ...d5, given the chance, and
White must always stop this.
Annotating the game in the October 1975
issue of 'CHESS ', Bronstein makes the
interesting observation that "Sax felt that
Black's central pawn position might prove
too rigid. Curiously, the effect is to make
 Bronstein − Move
65 by Move
White's position just as inelastic, as his
pieces are tied down to concentrating on his
d5-square, to prevent Black from playing ...
d6-d5".
QUESTION: So are you actually saying
Black is even better here?
ANSWER: No, I would not go that far, but
the position is balanced. The computer
claims an advantage for White, but I do not
really believe that.
14.♖ad1 ♖c8 15.♕b5 QUESTION: What is
that about?
ANSWER: Bronstein approved of this move,
which he describes as "a clever idea". The
white queen is heading to b3, from where it
attacks the e6-pawn, without being
vulnerable to the black rook on c8, as it
would be on c4.
♕c7 16.♕b3 ♕d7 17.♘b5
[ An important moment. The computer likes
17.♗xa7 , which was apparently Sax's
original intention. Bronstein gives as a
s a m p l e l i n e ♘xa7 18.♘xe5 ♕e8
19.♕xe6+ ♔h8 20.♘c4 ♖c6 "with little
advantage to either player", a line and
assessment with which the computer agrees.
Sax instead opts for a more cautious
alternative, but Bronstein thought this
was probably not quite so good. ]
17...b6
[ QUESTION: Isn't 17...a6 more natural?
ANSWER: Slightly, yes, and there seems
no objection to the move. Then the attempt
to reroute the knight towards the weakened
b6-square by 18.♘a3?! is well met by d5
.]
[ However, not 17...♘xe4? 18.♘xa7
A) 18...♖a8 is slightly better, but
19.♘xc6 ♕xc6 ( if 19...bxc6 20.♘xe5 )
20.♕xe6+ is still better for White;
B) 18...♘xa7 19.♘xe5! . ]
18.♘c3 ♖fd8 19.♖d3 ♗f8 20.♖fd1 ♕f7
Bronstein − Move by Move
QUESTION: Now what is happening? It
seems the play is fairly quiet.
ANSWER: Yes, both sides are regrouping.
White has doubled rooks against the d6-pawn,
but Black has been able to defend it easily.
21.♘h2?! QUESTION: What is the idea?
ANSWER: Well, as we will see, Sax intends
to bring this knight to g4 and exchange off
Black's king's knight, but it is not a very
impressive idea. On the other hand, it is not
easy to suggest a constructive plan for White.
His pieces are all dressed up with nowhere to
go, as it were – he has his major pieces
pressing against the theoretical weaknesses
on d6 and e6, but Black has them adequately
defended, and it is hard to see how White
can increase the pressure.
[ N o t e t o o t h a t 21.♘b5? is impossible
because of ♘xe4 . ]
21...♘a5 22.♕a2 ♘c4
[ A very natural move, but the computer
points out that Black missed the dynamic
chance 22...d5! .
QUESTION: But isn't that en prise?
ANSWER: Yes, but c2 will hang at the end.
Th i s i s a t y p i c a l S i ci l i an mo t i f , w i t h
Black's pressure down the c-file being a
recurring theme in such positions. After
23.exd5 exd5 24.♘xd5 (there is little
alternative, in view of the strength of the
black central pawns) ♘xd5 25.♖xd5 ♖xc2
White will have to play carefully to
maintain the balance. ]
23.♘g4 ♘xg4 24.hxg4 ♕g6
[ Bronstein notes that he rejected 24...♘xe3
25.♖xe3 ♕g6 , because he did not wish to
exchange off his strong knight on c4. It
does a very good job of shutting out the
white queen and shielding the central
pawns, as well as pressurizing the b2-
square. ]
25.♗c1
 Bronstein − Move
66 by Move
[ QUESTION: Why didn't he protect the g4-
pawn with 25.f3 ?
ANSWER: As Bronstein points out, that
w o u l d b e a mi s t a k e , b e c a u s e o f ♘xe3
26.♖xe3 d5! with the threat of 27...Bc5. ]
25...a6
[ Q U E S T I O N : B u t i s n ' t 25...♕xg4
just winning a pawn here?
ANSWER: No, White replies 26.♗xh6! .
( S t r a n g e l y , B r o n s t e i n g i v e s 26.♖g3
instead, but then after ♕h5 Black keeps his
extra pawn. )]
26.b3
[ QUESTION: He seems determined not to
defend his centre with 26.f3 .
ANSWER: Yes, Sax was obviously
reluctant to weaken the dark squares in this
way, until absolutely necessary. But he
should probably have done so here.
Bronstein then gives ♕f7 27.b3 ♘a5
28.♗e3 ♕c7 , which is still about equal. ]
26...♘a5 27.♗e3 ♘c6! This is the difference.
Black is able to exploit the dynamics, to
obtain a more active position than in the
previous note.
28.♕b2
[ The point is 28.♗xb6 ♘b4 29.♕b1
♘xd3 30.♗xd8 ♘xf2 (Bronstein),
( although the computer suggests that
30...♘f4 is even stronger, since White is
defenceless after 31.♗a5 ♕xg4 32.g3
♖xc3! 33.♗xc3 ♘e2+ 34.♔g2 ♘xc3
35.f3 ♕g5 36.♕c1 ♘xd1 , winning. )]
28...♘b4 29.♖3d2 QUESTION: Things
seems to have turned against White suddenly.
[ ANSWER: Yes, his position has become
extremely unpleasant. The computer
actually thinks he should jettison material
with 29.♗xb6 ♘xd3 30.♖xd3
, although Black remains clearly better after
♖d7 . ]
29...♖c6 Now White is in dreadful trouble.
Bronstein − Move by Move
Black will just double rooks on the c-file and
White is unable to avoid loss of a pawn,
since g4 is now hanging.
30.f3 ♖dc8 31.♘a2 d5!? QUESTION:
Why didn't he take the pawn?
[ ANSWER: Bronstein gave his last move
an exclamation mark, claiming that he
rejected 31...♖xc2 32.♕b1 "because I feel
I can get more out of the position".
Actually, his instinct was correct, but he
should have taken the pawn the other way,
as recommended by the computer. ]
[ I assume he rejected this because of
31...♘xc2 32.♗xb6 , but now d5!
is tremendously strong, 33.-- ( the nasty
point being that 33.exd5? opens the black
queen's line of defence of the knight at c2,
thus losing a whole piece after ♖xb6 .;
33.a5 is relatively best, but after d4
White's position is hopeless, as his queen
is boxed in on a2 and his bishop is shut out
of play on b6. Black will infiltrate on the
kingside dark squares with ...Qg5,
combined with a timely ...d3, ...Rxb6 and
...Bc5+. )]
32.♘xb4
[ White's last chance, not mentioned by
Bronstein, was 32.c4! dxc4 ( 32...dxe4?
33.♘xb4 ♗xb4 34.♖d8+ ♔h7 35.♕xe5
l e t s W h i t e r i g h t b a c k i n t o t h e g a m e)
33.bxc4 ♖xc4 ( or 33...♘xa2 34.♕xa2
♖xc4 35.♗xb6 ) 34.♘xb4 ♖xb4 35.♕xe5
when White is still on the board. Instead,
after the text, he goes down without
much further resistance. ]
32...♗xb4 33.♖e2 d4
[ Perhaps predictably, the computer prefers
to grab a pawn by 33...♗c3 34.♕a3 dxe4
, when Black should win easily enough,
but Bronstein commented that "I prefer to
eliminate with one pawn move all four
hostile pieces". ]
 Bronstein − Move
67 by Move
34.♗f2 ♗e7
[ 34...h5! is more energetic. Bronstein's
move intends ...Bg5, but is arguably less
incisive. However, a relevant factor here
is that Sax was very short of time, having
only two minutes left for six moves. In
such a situation, quiet moves like 34...Be7
are much more unpleasant for the defender
to face than forcing lines. ]
35.♖d3 b5 36.axb5 axb5 37.f4?
With his flag hanging, Sax collapses, making
a desperate and suicidal bid to undermine the
cramping black d4-pawn.
[ Instead, he should have played something
solid, such as 37.♖d1 , when Black still has
to demonstrate how he will win. ]
37...♕xg4 38.♖f3 ♖c3
[ White is in a comic state of helplessness,
since 38...♖c3 threatens 39...Rxf3, and
39.♖xc3 dxc3 drops the other rook.
An interesting positional game by
Bronstein, who showed how the apparently
weak black central pawns can actually be
an asset in certain circumstances. His
finishing from move 31 onwards was not
as incisive as it might have been, but the
computer always tends to make humans
look silly in such positions. Bronstein's
handling of the position, especially given
his opponent's time-trouble, was typically
pragmatic for the human player, even
though he missed objectively stronger
lines. ]
0-1
Bronstein − Move by Move
26 B99
Bronstein,D
Browne,W
Reykjavik (Game 17) 1990
[Steve Giddins]
1.e4 c5 2.♘f3 d6 3.d4 cxd4 4.♘xd4 ♘f6
5.♘c3 a6 The Najdorf Variation, which was
Walter Browne's calling-card against 1 e4.
The American GM never played anything
else as Black, and scored a huge number of
points in the line. But, inevitably, he
occasionally came unstuck, through walking
into his opponent's preparation.
6.♗g5 The sharpest response, especially for
those days.
QUESTION: It was quite unusual for
Bronstein to go down long theoretical lines,
wasn't it?
ANSWER: Yes it was. His restless creativity
usually led him in less-explored directions.
But there is a story attached to this game. On
YouTube, you can find an interesting ten-
minute video, in which Yasser Seirawan
talks about this game (the exact link is https:/
/ w w w . y o u t u b e . c o m /
watch?v=SAymAaV92qk).
Seirawan was playing in the event, and
records that Bronstein phoned him up before
the game, invited hi m to his room, and
explained that he was playing Browne that
afternoon. Bronstein then went on to say that
he wished to take on the Najdorf, "as a
tribute to your countryman", and showed
Seirawan a complicated middlegame position
that he was aiming to reach. Seirawan
counselled against this, pointing out that
Browne knew the Najdorf backwards, but
Bronstein just dismissed such objections,
continually pointing at the position and
saying "But look at the possibilities!".
Later that afternoon, Seirawan walked over
 Bronstein − Move
68 by Move
to Bronstein's board, and saw the very
position he had been shown in the latter's
hotel room that morning. Unfortunately, in
the video, he does not give the exact position,
but further on, I have pointed out which
position I suspect it was.
e6 7.f4 ♗e7
[ 7...♕b6 is the so-called Poisoned Pawn,
which is the most popular variation at the
time of writing, but Browne's choice was
always the main line, and was his own
regular preference. ]
8.♕f3 ♕c7 9.0-0-0 ♘bd7 10.g4
[ 10.♗d3 is the alternative, played in the
Spassky-Fischer match of 1972, but then
Browne himself was largely responsible
for developing the reply h6! which is
regarded as best for Black, and gives him
a decent position, according to theory. As
a result, attention shifted back to th e
older move in the game. ]
10...b5 11.♗xf6 ♘xf6 QUESTION: Is this
forced?
ANSWER: Almost.
[ 11...♗xf6?! is known to be good for
W h i t e a f t e r t h e s a c r i f i c e 12.♗xb5
(the point being that d6 is undefended), ]
[ whilst Fischer's old hobby-horse 11...gxf6
has never actually been refuted, to the best
of my knowledge, but is not nowadays
regarded as terribly good. ]
12.g5 ♘d7 13.f5! The most aggressive.
QUESTION: But losing a pawn with check?
ANSWER: Yes. White gambits a pawn, to
open the g-file and lure the enemy bishop to
an exposed square.
♗xg5+
[ Again, nowadays accepted as best,
although a great deal of research was spent
establishing that the alternatives 13...♘e5 ]
[ and 13...♘c5 were less good. ]
14.♔b1 ♘e5 15.♕h5 ♕d8?! I am indebted
Bronstein − Move by Move
to Grandmaster Matthew Sadler for his
insights into this game, as he plays the
Najdorf himself and is an expert on this
whole variation. It seems that this is where
things start to go wrong for Black.
[ These variations are bewilderingly
complex, and theory took a long time to
establish the best lines for both sides, but
nowadays, 15...♕e7 is regarded as slightly
more accurate.
QUESTION: What is the difference?
ANSWER: Well, the detailed variations are
far too long and complex to delve into here,
but the main general point is that on e7,
the queen defends e6 and f7, whereas from
d8 it does not. That makes a difference in
various lines, but none but Najdorf buffs
need worry their heads over exactly what
difference. ]
16.♖g1
[ M a t t h e w t e l l s m e t h a t 16.h4!
is now regarded as stronger, since the h-
pawn can become quite a dangerous
attacker in various lines. ]
16...h6?! Matthew's comment was that this
was the biggest error of all. Black's last
move is simply too weakening.
QUESTION: So what should Black do about
his bishop?
[ ANSWER: He should retreat at once with
16...♗f6 . Then after 17.fxe6 both 0-0
( and 17...g6 have been extensively tested,
with results that are not too bad for Black. )]
17.fxe6 g6 18.exf7+ ♔xf7 19.♕e2 ♔g7
20.h4!
[ Up to this point, Black had been smashing
out his moves at blitz speed, as Browne
was still in his book. The previous year, he
had reached this position, against the
young American GM, Patrick Wolff. That
game had continued 20.♘d5 ♖f8 21.♕g2
♖a7 22.♕g3 ♔h7 and Black was OK and
 Bronstein − Move
69 by Move
eventually won, P.Wolff-W.Browne,
Philadelphia 1989. ]
QUESTION: So was Bronstein's move a
novelty?
ANSWER: Not exactly. The database has
one earlier game where 20 h4! was played,
but it was quite an obscure game, and I am
not sure if Bronstein knew of it. One must
remember that there were no databases in
those days, and information travelled much
more slowly. I suspect Bronstein probably
found 20 h4! himself. The move is natural
enough, offering a pawn to open additional
lines on the kingside, and to gain access to
the weakened g6-square.
20...♗xh4!? Played after a whole hour's
thought, according to Seirawan, which
would suggest that even the Najdorf addict
Browne was unaware of the previous game
with 20 h4. I suspect this was the position
Seirawan had seen on the board in
Bronstein's hotel room that morning. He
describes the position as one where White
was a pawn down, and where he (Seirawan)
could not see any obvious threats, so this
must be it.
QUESTION: But Bronstein thought it was
full of possibilities?
ANSWER: Yes, and the computer agrees
with him, even assessing it as winning for
White.
QUESTION: Does Black have anything
better than taking the pawn?
[ A N S W E R : F r i t z s u g g e s t s 20...b4
as slightly better, but even so, after 21.♘a4
( 21.♘d5 looks more obvious, but White
wants to keep the d-file open) 21...♗e7
22.♘c5! the white knights irrupt into the
enemy position, with great power. This
looks pretty awful for Black. ]
21.♘f5+ ♔h7
[ The previous game with 20 h4! had
Bronstein − Move by Move
continued 21...♗xf5 22.exf5 ♖e8 23.♗h3
♗g5 24.fxg6 ♘c4 25.♕d3 ♕e7 26.♖gf1
, R.Tischbierek-M.Schurade, Nordhausen
1986, with what should have been a
decisive advantage to White, although he
actually went on to lose, after falling into
a mating net in an ending where he was
an exchange and several pawns up.
QUESTION: But you think the position
here is just winning for White?
ANSWER: Well, certainly much better.
QUESTION: Why?
ANSWER: The key factor is the exposed
black king. The presence of opposite-
coloured bishops makes things much worse
for Black, as his light squares are terribly
weak, and White effectively has an extra
piece with which to attack them. ]
22.♖xd6 QUESTION: Is this any better for
Black?
ANSWER: Not much. White has regained his
pawn, and has much the more active pieces,
whilst the black king remains terribly weak.
The opening has been a disaster for Black.
QUESTION: But how come? After all,
Browne was following his theory until a few
moves ago, yet after a single new move, he
seems to be almost lost.
ANSWER: That is always the risk with such
sharp variations, especially back in the pre-
computer era. Nowadays, one only has to put
the position after 19...Kg7 on any decent
chess engine, and it immediately screams
that 20 h4 yields a clear advantage. But in
those days, players had to analyse everything
by themselves, and there are just so many
possibilities in these complicated variations
that they would inevitably miss things now
and then.
♕f8 23.♕h2 ♗xf5
[ Forced, since two black minor pieces are
hanging, and the attempt to defend them
 Bronstein − Move
70 by Move
both with 23...♗f6 runs into 24.♖xf6!
♕xf6 25.♘d5 ( 25.♕xh6+? ♔g8
is much less clear) 25...♕f8 and now there
is a beautiful finish.
EXERCISE: Can you find a smart mate
here?
ANSWER: 26.♕xh6+! ♕xh6
( actually, Black can delay the mate a few
moves with 26...♔g8 , but it is still mate in
five: 27.♖xg6+ ♔f7 28.♖g7+ ♔e8
29.♘c7+ ♔d8 30.♕g5+ and mate next
move ) 27.♘f6# is a beautiful mate. ]
24.♕xe5 ♕e7
[ The exchange of queens brings no relief
to Black's position, but the alternative
24...♖e8 is no better after 25.♕d5 ♗c8
26.♗d3 , with Rf1 to follow. ]
25.♕xe7+ ♗xe7 QUESTION: Surely the
queens coming off must help Black, since
his exposed king should now be less of an
issue?
ANSWER: One would think so, but his king
still proves a millstone round his neck, as the
white rooks are so much more active and
Black's bishops lack stable outposts. This is
an instructive example of the white initiative
extending into the endgame. We have seen
several examples of this phenomenon already
in this book.
26.♖c6 ♖hc8!
[ This tactical resource is the only way to
avoid immediate loss of material, since
after 26...♗d7 the bishops run into lethal
trouble: 27.♖c7 ♖hd8 28.♘d5 ♗g5 ( or
28...♗d6 29.♖xd7+ ♖xd7 30.♘f6+ )
29.♗h3 and Black is losing a piece,
( but not 29.♘b6?? ♗e3 . )]
[ Similarly, 26...♗c8 27.♖c7 ♖e8 28.♘d5
is no better. ]
27.♖b6 ♖xc3? QUESTION: What is this?
Desperation?
ANSWER: Exactly.
Bronstein − Move by Move
[ Retreating the bishop is not a lot better
than in the previous note: for example,
27...♗d7 28.♖b7 ♖d8 29.♗h3 ♗xh3
30.♖xe7+ ♔h8 31.♖xg6 and it is all
over. ]
[ QUESTION: But what about 27...♗c5
, hitting both white rooks?
ANSWER: That is much the best try, and
the only way to stay on the board. After
28.♖b7+ ♔g8 29.♖h1 ♗e6 30.♖xh6 ♗f7
Black is a pawn down, but he is holding
the remains of his position together and
his bishops offer some defensive chances in
the ending. ]
28.exf5!
[ M u c h s t r o n g e r t h a n 28.bxc3 ♗c5
with excellent drawing chances. ]
28...♖e3 29.♗d3 QUESTION: Opposite-
coloured bishops in the ending are supposed
to be a drawing factor, aren't they?
ANSWER: Pure opposite-coloured bishop
endings have a high drawing factor, yes, but
when there are other pieces on the board,
this effect is much reduced. Here, we still
have something more akin to a middlegame,
with the black king still subject to attack.
The opposite-coloured bishops actually
strengthen that attack, by enabling White to
bring an extra piece to bear on the light
squares.
♗c5
[ 29...g5 looks as though it drops a piece
after 30.f6+ , but, paradoxically, this was
Black's best chance. After ♖xd3 31.cxd3
( 31.fxe7 ♖e3 is nothing for White)
31...♗c5 32.♖b7+ ♔g6 33.♖f1
White's extra exchange should win, but
Black's passed kingside pawns offer some
hopes of defending the position. ]
30.♖bxg6 Now it is much simpler, as the
black king is still in serious trouble.
♖ae8 31.a4 bxa4 32.f6! A neat conclusion.
 Bronstein − Move
71 by Move

White brings his light-squared bishop into championship level, surely?

the attack. ANSWER: Yes, indeed, but Botvinnik was
♖xd3 always fond of it, ever since his youth, and it
[ The only other defence to the threat of 33
featured extensively in this particular match
R g 7 + a n d 3 4 R h 7 m a t e w a s 32...♖e1+
against Bronstein, with both players using it.
, but after 33.♖xe1 ♖xe1+ 34.♔a2 ♔h8 Botvinnik's choice of it in this game is
35.♖xh6+ mate is forced: ♔g8 36.♗c4+ doubly interesting, though, in view of the
♔f8 37.♖h8# . ] match situation. At this point, the match was
33.♖g7+! ♔h8 34.♖h1! all square at 10½-10½, with Bronstein
[ Black can avoid the mate with 34.♖h1 having just won Game 21 as Black, to
♗e3 , b u t a f t e r 35.cxd3 he is clearly
equalize. Under the rules of the day,
losing. Botvinnik needed only to draw the match
QUESTION: So, going back to Seirawan's 12-12, in order to retain his title, so one
video, it seems that Bronstein's idea of might have expected him to choose a more
"showing respect" for Browne and his solid defence.
Najdorf was to smash him off the board! QUESTION: So why didn't he?
ANSWER: Yes! Bronstein was like that, ANSWER: Paradoxically, I think he regarded
though. I don't believe his pre-game the Dutch Stonewall as being solid. He had
comment was meant at all ironically – he always had great faith in it, ever since his
will have regarded it as a tribute to his youth, and had used it against such titans as
opponent to take him on in his favourite Capablanca, Keres and Reshevsky, without
opening. Of course, having a nice new(ish) losing, so I think in his mind, it was as solid
idea like 20 h4! helps, but even so, it wasand reliable a line as any.
typical Bronstein to put non-sporting 3.g3 ♘f6 4.♗g2 ♗e7 5.♘c3 0-0 6.e3
considerations ahead of pure pragmatism. This is a little unusual, but once again,
Most players would have agreed with psychological games are being played here.
Seirawan that it made no sense to take Botvinnik himself had played this move as
Browne on in his favourite opening, but White in the first game of the match, when
Bronstein was never keen on such Bronstein had chosen the Dutch. As the
materialistic and pragmatic latter puts it in 'David Bronstein: Chess
considerations. ] Improviser', "Botvinnik plays 'my' Dutch
1-0 Defence and I play his system with 6 e3".
d5
[ The aforementioned game one had seen
27 A91 Bronstein prefer 6...d6 here, and after the
Bronstein,D further moves 7.♘ge2 c6 8.0-0 e5 9.d5
Botvinnik,M ♕e8 10.e4 ♕h5 11.exf5 ♗xf5
22nd matchgame, Moscow WCh. (Game 18) Black was doing fine. Botvinnik remains
[Steve Giddins] faithful to his beloved Stonewall set-up, but
I am not sure it is really best in this
1.d4 e6 2.c4 f5 QUESTION: The Dutch position. ]
Defence is a pretty rare weapon at the world QUESTION: Why is that?

Bronstein − Move by Move

 Bronstein − Move
72 by Move
ANSWER: With White's knight on e2, I like
the idea of getting a black pawn to e5 as
soon as possible, so that the pawn
'dominates' the knight, stopping it jumping to
d4 or f4. On the other hand, it seems to me
that the knight is quite well-placed on e2
against the Stonewall, as it can transfer to f4
and d3, eyeing up the key e5-square, and it
also leaves the white f-pawn free to advance
to f3, supporting the break e4. On the other
hand, Botvinnik had forgotten more about
the Dutch than I will ever know, and he had
also played 6...d5 in Game 16 of the same
match, so he clearly believed in it.
7.♘ge2 c6 8.b3 ♘e4
[ In Game 16, Botvinnik had played 8...♗d6
9.0-0 ♕e7 , a typical regrouping in such
positions, and had achieved a satisfactory
position. The game was eventually drawn
after a gruelling fight, with Black close to
a win in the endgame.
QUESTION: So why did he deviate now?
ANSWER: It is not clear. Possibly he
feared an improvement, although it is not
really the type of position where a strong
novelty is likely – the position is about
patient manoeuvring. The fact that
Botv inn ik dev iates f rom a succes sfu l
previous game says quite a lot about his
character and objectivity, I suspect. Even
when an opening had gone well for him, he
would analyse the game closely and still
look for improvements. ]
9.0-0 ♘d7 10.♗b2 ♘df6 QUESTION: So
how do you assess the position?
ANSWER: It is a typical Stonewall, where
White h as a s light advantage, p erhaps ,
thanks to the weakness of e5, but where
Black is very solid.
QUESTION: I thought the Stonewall was
supposed to be bad, so how come White
doesn't have the advantage?
Bronstein − Move by Move
ANSWER: I think the popular view of the
Stonewall is just incorrect. The opening is
not bad at all, but it does require careful
handling and deep understanding.
Unfortunately, there are many relatively
weak players, who have used the line over
the years, usually hoping they can simply
give mate by ...Qe8-h5, etc. The trouble with
such a one-dimensional approach is that it
does not really work against decent white
play, and there have been numerous games
where Black has suffered a fiasco in such
positions.
We have all seen many examples of
nightmare bad Stonewall bishops, etc, and
the opening has gained a bad reputation as a
result. But when it is handled by a seriously
strong GM, who knows what he is doing, the
positions are perfectly playable and also
highly interesting. For an excellent
explanation of some of the fine points of the
opening, I can recommend the section in the
2012 book, 'Study Chess with Matthew
Sadler'.
11.♕d3
[ Kasparov, quoting Boleslavsky, gives
11.f3 as more accurate, with the idea of
♘xc3 ( or 11...♘d6 12.♕d3 and again, a
quick e4 ) 12.♘xc3 followed by e4. This
supports my contention that the Stonewall
may not be the most accurate reaction to
White's e3 and Nge2 set-up. ]
11...g5!? QUESTION: Wow! He is really
going for it, isn't he?
ANSWER: Funnily enough, this is not quite
the attacking gesture it looks. Black has a
more subtle, positional idea in mind – he
wants to play ...g4, so as to secure control of
the e4-square and stop White organizing f3
and e4.
QUESTION: So is it any good?
[ ANSWER: It is interesting, but Kasparov
 Bronstein − Move
73 by Move
recommends instead a nice way of
exploiting White's slight inaccuracy on the
previous move: 11...♗d7! .
QU ES TI O N: H u h , t h at d o es n o t lo o k
terribly impressive or striking?
ANSWER: Maybe not, but there is a subtle
point to it. The idea is to meet White's
intended 12.f3 with ♘xc3 13.♘xc3 ♗e8
14.e4 fxe4 15.fxe4 ♗g6 . Now the pin on
the e4-pawn is awkward for White, and
after 16.♕e3 there is the further clever
point dxc4 17.bxc4?! ♘g4! and suddenly,
White's position is a bit hollow in the centre.
His queen is attacked, he must guard d4,
and ...Bg5 and/or ...e5 are in the air .
Black would have good counterplay. ]
12.cxd5
[ White might have done better to dispense
with this exchange and to play the
immediate 12.f3 , with a continuation
similar to the game. ]
12...exd5 13.f3 ♘xc3 14.♗xc3 g4
This is the point of Black's idea, begun at
move 11. He keeps fighting to undermine
White's control of e4.
15.fxg4 ♘xg4 QUESTION: So how do you
assess this position?
ANSWER: The computer likes White a lot,
but I think his advantage is only relatively
small, an assessment Kasparov confirms.
White has weaknesses on e3 and e4, Black
one on f5 and the draughtier king.
16.♗h3
[ The computer prefers 16.♘f4 , but that
would allow ♘f6 and a later ...Ne4, which
Bronstein instead hurries to prevent. ]
16...♘h6 Now the knight is much less well-
placed on the edge of the board, and tied to
defending f5.
17.♘f4 ♗d6 18.b4! QUESTION: What is
this, a minority attack?
ANSWER: Just so. White aims to play b5
Bronstein − Move by Move
and create a weakness in the black queenside.
a6 19.a4 ♕e7 20.♖ab1
[ The immediate 20.b5 was possible, but
Bronstein prefers to make another
preparatory move, giving Black the chance
to take radical measures on the queenside. ]
20...b5!? QUESTION: That is certainly
radical! But is the cure worse than the
disease?
ANSWER: It is one of the standard ways of
meeting the mino rity attack , bu t is no t
always open to Black. Basically, for this to
work, he needs to be able to prevent White
from playing e4 (for a classic example of
what can happen if he does not, see the final
game of the Karpov-Spassky, Candidates
Semi-final of 1974). Here, Black has a firm
grip on e4 (at least, at the moment), so in
principle, conditions ought to be right for ...
b5. However, it is committal, and one of the
other key aspects of successful application of
s u ch a s t r u c t u r e , n amel y t h e ab i l i t y t o
transfer a black knight to c4, is not present
here.
[ Kasparov indicates that he would have
preferred 20...♗d7 , with the idea of 21.b5
axb5 22.axb5 ♘g4 , and the possession of
the open a-file gives Black reasonable
play. ]
21.♗g2 ♘g4 Seizing the chance to transfer
the knight to a better square.
22.♗d2 ♘f6 23.♖b2 QUESTION: What is
the idea of this?
ANSWER: White has the plan of seizing the
a-file, by playing Ra1, to threaten axb5.
QUESTION: And how is the assessment
now?
ANSWER: Black is holding the balance. His
knight is now on a much better square than
before, with the further possibility of coming
to e4 and thence possibly to d6, always a
good square for a black knight in these
 Bronstein − Move
74 by Move
minority attack structures.
♗d7!?
[ Bronstein criticized this as allowing White
a fav o u r ab l e regroup i n g , a n d i n s t e a d
recommended 23...♘e4 .
QUESTION: Why is that better?
ANSWER: Well, Bronstein's idea was that
White would either have to retreat the
bishop to e1, interfering with the transfer
of his rooks to the a-file, or "lose his
important bishop".
QUESTION: But that looks rather a
crummy bishop, so isn't White happy to
trade it for the black knight?
ANSWER: Exactly! That is what Kasparov
also says – he recommends 24.♖a1 ♘xd2
25.♕xd2 , followed by Nd3 and Bf3 when
Black still lacks counterplay.
QUESTION: A strange misjudgement by
Bronstein, isn't it?
ANSWER: Well, maybe, but he probably
thought that the white queen's bishop,
although rather passive, is doing an
important job of defending the pawns on e3
and b4 and that it would be missed after
its exchange. There is some truth in that,
and it is not clear how much advantage
White really has, even in Kasparov's line.
One other factor is that Bronstein's
attitude in his notes may have been
influenced by what happens later in the
game. As we will see, the supposedly bad
queen's bishop plays a decisive role in the
final attack. Even world class players can
find it difficult to shake off such
impressions. ]
24.♖a1 ♘e4 25.♗e1 Now White is able to
preserve the bishop, at no inconvenience to
himself, so it certainly makes sense to do so.
♖fe8 26.♕b3 ♔h8?!
[ 26...♗xf4!? was also an option here,
removing the strong white knight.
Bronstein − Move by Move
QUESTION: But then his dark squares
look like a bad accident!
ANSWER: True, but it is not that easy for
White to get at them, because his central
pawns are fixed on dark squares.
QUESTION: So why didn't Botvinnik play
this?
ANSWER: It is not clear. Bronstein says
that he probably feared a later Bh4 (after
27.gxf4 ), but Kasparov suggests that
Botvinnik may instead simply have thought
he was holding the game position, and had
no need to make this exchange. ]
27.♖ba2 ♕f8?! Missing the last chance to
take on f4.
28.♘d3! The knight moves away, putting
paid to any thoughts of ...Bxf4, and also
defends the b4-pawn while taking aim at c5
and e5. Now a possible Ne5 must always be
reckoned with, hoping to open up the long
diagonal, if Black should capture twice on e5.
♖ab8?! Neither Kasparov nor Bronstein
comment on this abandonment of the a-file,
but it seems to help White. On the other hand,
it is not easy to find a decent alternative.
[ T h e c o m p u t e r s u g g e s t s 28...♕h6
, trying to exploit White's last move by
attacking e3, but even so, after 29.♘c5
Black is still suffering. ]
29.axb5 axb5 30.♖a7 Adding another
advantage to White's collection – he has the
seventh rank. Now there is no doubt about
his advantage.
QUESTION: So where did Black go wrong?
He seemed OK half a dozen moves ago.
ANSWER: Indeed. It is hard to pinpoint what
has gone wrong, but I suspect he should
have traded on f4 when he had the chance, as
discussed above.
♖e7 31.♘e5! Things are now really starting
to come together for White. His knight
irrupts into the enemy position, exploiting
 Bronstein − Move
75 by Move
the classic Stonewall weakness on e5, and
Black's game is now extremely hard to
defend.
♗e8
[ 31...♗xe5 32.dxe5 looks terribly
dangerous, but now a blow of fearful
strength demolishes the black position. ]
[ EXERCISE: Can you find a good
positional move after 31...♗e8 ?
Looking at the position, one can see that
the white dark-squared bishop on e1
would really like to get into the game, so
as t o ex p l o i t B la ck 's n u me r o u s d ar k -
square weaknesses. So, Bronstein played: ]
ANSWER: 32.g4!! A lovely positional blow.
With White's pieces mainly infiltrating down
the queenside, Bronstein suddenly opens a
second front on the other wing. Now his
bishop on e1 enters the game with decisive
effect.
fxg4? With time-trouble adding to his other
woes, Botvinnik collapses.
[ As Kasparov points out, the computer
suggests Black is still hanging on after the
co ld - b l o od ed 32...♗xe5 33.dxe5 ♖g7
34.♖xg7 ♕xg7 35.gxf5 ♕xe5 36.♖a7
♔g8 , ]
[ w h i l s t t h e i m m e d i a t e 32...♖g7
is also more tenacious than the text. ]
33.♗xe4 dxe4 34.♗h4 ♖xe5?
Another mistake, which soon ends the game.
[ Once again, the computer shows the
toughest defence as 34...♖xa7 35.♖xa7
, and now two queen moves are possible:
A) 35...♕h6 is one, after which 36.♗g3
♖d8 37.♗f4 ♕f6 38.♖f7!
( much stronger than Kasparov's 38.♕c2 )
38...♗xf7 39.♘xf7+ ♔g7 40.♘xd6
is decisive.;
B) Kasparov gives the other queen move,
35...♕f5 , as best of all.
B1) Bronstein also considers this, and
Bronstein − Move by Move
points out the variation 36.♖f7 ♗xf7
37.♘xf7+ ♔g7 38.♘xd6 ♕f3 39.♗g3
♖a8 , which the computer confirms is
only a draw.;
B2) Instead, Bronstein recommended
36.♗g3 "when there is no defence
against 37 Rf7". Kasparov queries this,
c i t i n g t he l i n e ♖d8 37.♖f7 ♕g5
38.♖f4 ( but here, Fritz finds the
improvement 38.♗f4! followed by 39
Qe6, which wins at once; so Bronstein's
3 6 B g 3 s e e m s t o w i n i n t h i s l i n e)
38...♕g8 .;
B3) Nevertheless, Kasparov is right that
the best response of all to 35...Qf5 is
the typical computer move, 36.♕a2!!
when the queen suddenly threatens to
penetrate down the a-file in various
lines. For example, ♗f8 37.♘f7+ ♗xf7
38.♖xf7 ♕d5 39.♕a7 winning.;
B4) 36.--;
C) 35...-- ]
35.dxe5 ♗xe5 36.♖f1 Now, on the other
hand, the game ends prosaically.
♕g8
[ QUESTION: Is 36...♕h6 any better?
ANSWER: Unfortunately not. Then
37.♖f5! leaves the black bishop without a
safe square on the long diagonal, as ♗g7
38.♖xg7! ♕xg7 ( 38...♔xg7 39.♕b2+
♔g8 40.♕a2+ ♔g7 41.♕a7+ wins )
39.♗f6 decides. ]
37.♗g3! Fatally luring the bishop off the long
diagonal.
♗g7
[ 37...♗xg3 38.♕c3+ mates, ]
[ as does 37...♕xb3 38.♖f8+ ♕g8
39.♗xe5# . ]
38.♕xg8+
[ After 38.♕xg8+ ♔xg8 , 39.♗xb8 wins.
A beautiful game by Bronstein, and one
which set up a thrilling finish to the match.
 Bronstein − Move
76 by Move

Botvinnik was left a game down, with two desirable e4 advance.

to play, and effectively had to win with [ B l a c k ' s m a i n a l t e r n a t i v e i s 4...c6
White in Game 23. He did so, in a very , transposing back into the Slav Defence.
famous two-bishop endgame, which you That is quite a good, solid choice, but the
can find with detailed notes in my book Slav is a different opening from the QGA,
'The Greatest Chess Endgames Ever and, if he really wanted to play it, Black
Played'. ] could have done so at move 2.
1-0 QUESTION: But he has avoided some
lines, such as the Exchange Variation,
hasn't he?
28 D24 ANSWER: True. Going into the Slav in
Bronstein,D this way is a perfectly reasonable option,
Korchnoi,V and has some merits in move order terms,
Moscow Zonal (Game 19) 1964 as you say, but it still involves knowing
[Steve Giddins] two different openings. Not everyone wants
to learn the main line Slav, just to counter
1.d4 d5 2.c4 dxc4 3.♘f3 ♘f6 4.♘c3 a rarely-seen gambit like 4 Nc3. ]
QUESTION: That is quite unusual, isn't it? 5.e4 b5 QUESTION: So Black has won a
ANSWER: Yes. It is the introduction to a pawn?
sharp gambit continuation, as Black can now ANSWER: Yes, but White has taken control
play to keep his extra pawn. of the centre and has more space. This is
QUESTION: Is the line any good? going to result in a sharp battle, where White
ANSWER: It is quite dangerous, actually. It n e e d s t o p l a y v i g o r o u s l y , i n o r d e r t o
had a spell of popularity back in the 1990s, capitalize on his trumps. If he dawdles, and
being employed by Kramnik, for example. allows Black to complete his development
He is not a player known for his love of with ...Bb7, ...e6, ...Be7 and ...0-0, for
gambit play, so I think his espousal of the example, then White will just find himself a
line can be taken as evidence that it has pawn down for nothing.
greater merit than just being a piece of 6.e5 6...b4 was threatened, so this is the
speculation. obvious and consistent follow-up. Black's
[ The usual move is 4.e3 , of course, ] knight is driven from its customary position,
[ whilst 4.♕a4+ is also possible. ] defending the kingside.
4...a6 This is the point. With the white knight ♘d5 7.a4 ♘xc3 This was the main line at
on c3, Black can now play ...a6 and ...b5, the time, and remains an important variation
trying to hang on to the extra pawn, which to this day.
would not have succeeded a move earlier. [ However, despite his hammering in the
QUESTION: Does he have any alternative? present game, Korchnoi clearly did not
ANSWER: His choice is quite restricted, as lose faith in Black's position. Some thirty-
if he does not hang on to the pawn, the o d d y e a r s la t e r , h e d ef e n d e d t h e l i n e
danger is that White will just play e4 and against Kramnik, with the new continuation
Bxc4, establishing a more favourable version 7...e6 8.axb5 ♘b6 .
of the main lines, where he has achieved the QUESTION: What is the point? Black

Bronstein − Move by Move

 Bronstein − Move
77 by Move
seems to have lost back his extra pawn.
ANSWER: Yes, but he hangs on to the
pawn on c4, which prevents White's
king's bishop from developing to an
aggressive position on d3. This line had a
brief vogue of popularity at the time, and
is still regarded as probably Black's best
response to 4 Nc3. ]
8.bxc3 ♕d5 QUESTION: Wow! That looks
a bit strange. Black has no pieces developed
at all, yet he brings out his queen. What
happened to basic opening principles?
ANSWER: Well, such principles are
important, of course, but specifics always
have to predominate. White had a direct
threat of 9 axb5, exploiting the pin against
the undefended rook on a8, so Black needed
to meet this. He has a very limited choice of
ways to do so.
[ The only real alternative was 8...♗b7
, but this allows White a classic blockading
sacrifice with 9.e6 , offering another pawn,
with the aim of gumming up Black's
development. Three years later, Bronstein
had a game with this very line, which
continued fxe6 10.♘g5 ♕d5 11.♕g4
♕f5 12.♕g3 and now Black felt obliged to
return one of his extra pawns to douse the
white initiative. After e5 13.♕xe5 ♕xe5+
14.dxe5 ♘d7? ( 14...♗d5! holds the
balance ) 15.♘e6 ♖c8 16.axb5 axb5
17.♖a7 White had a large advantage, and
went on to win in D.Bronstein-I.Chikovani,
USSR Team Championship 1967. ]
9.g3 With the black queen exposed on d5, it
makes sense to fianchetto White's bishop,
setting up a threat of Bg2 and then Nh4.
♗b7 10.♗g2 ♕d7 QUESTION: This just
loses a tempo, though, doesn't it?
ANSWER: Yes, but Black had to meet the
threat of 11 Nh4. Essentially, 8...Qd5 was a
temporary measure – Black wanted to defend
Bronstein − Move by Move
the a8-rook, whilst also stopping the thrust
e5-e6. Since 8...Bb7 allows 9 e6, as we have
seen, he sacrifices a tempo, to play ...Qd5,
then ...Bb7 and ...Qd7.
QUESTION: But I don't like sacrificing
tempi in the opening!
ANSWER: Maybe not, but you do like
pawns! As the Roman poet Virgil observed,
there is no such thing as a free lunch, and
that applies on the chessboard as well. Black
has nicked a pawn and hung on to it, so he
must expect to have to suffer a bit. If it were
possible to win the pawn without any
inconvenience at all, then nobody would
offer the gambit with 4 Nc3, of course.
11.♗a3! Another accurate move. Before
Black has the chance to play e6, White puts
his bishop on the a3-f8 diagonal, thus
ensuring that ...e6 will involve the loss of
castling rights for Black.
e6
[ QUESTION: Is 11...g6 possible? It looks
more attractive than losing castling rights.
ANSWER: It is possible, but after 12.0-0
♗g7 13.♖e1 Black once again has to
reckon with the threat of e6, shattering his
pawn structure.
Q U E S T I O N : O K , b u t t h e n I p l a y ♗d5
stopping it.
A N S W E R : Y e s , a n d n o w 14.♘g5!
is unpleasant for you. Again, e6 is a threat,
and e6 leaves you unable to castle and with
serious weaknesses on the kingside dark
squares. I continue 15.♘e4 and you have
some very unpleasant problems to solve.
QUESTION: So was there no other
alternative? ]
[ ANSWER: The computer suggests
11...♗d5 12.0-0 ♘c6 , developing the
queenside, whilst postponing a decision
about the other wing. That makes some
sense, but I do not think it really solves the
 Bronstein − Move
78 by Move
issue – it merely postpones it. After simply
13.♖e1 g6 14.♘g5 , I am not convinced
that much has changed. ]
12.♗xf8 ♔xf8 13.0-0 QUESTION: So what
is your assessment of this position?
ANSWER: I rather like White's game. He
has excellent compensation for the pawn, in
the shape of more space, better development,
and some potentially serious dark-square
weaknesses in the black camp, notably c5.
QUESTION: So are you saying that the
whole variation is just bad for Black?
ANSWER: I would not go that far. Firstly,
he has the later Korchnoi idea 7...e6 8 axb5
Nb6, which I think is better than the game
continuation. And, secondly, even the game
position may not be that bad for him, but I
think it takes a certain kind of player to
defend such positions. Korchnoi was always
a notorious pawn-grabber, and won many
games by nicking hot-looking pawns and
somehow defending the threats, but it is not
a strategy at which many players excel. And,
as we will see, even he could come badly
unstuck on occasion. So I would not
recommend this line to Black as a practical
proposition, even if the computer claims that
he is objectively OK.
g6!? QUESTION: Goodness me! You were
talking about dark-square weaknesses, and
now he play s that! Surely this mu st be
terrible?
ANSWER: It is certainly very ugly, but it is
not easy to suggest a good alternative. Black
needs to find a way to connect his rooks, and
there is no other obvious means of doing so.
14.♘h4?! I am a little surprised by this, I
must admit.
[ 14.♘g5 is much more natural, planning to
bring the knight to e4 and exploit the dark
squares. I think Black's game would be
pretty difficult then.
Bronstein − Move by Move
QUESTION: So why do you think
Bronstein didn't play it?
ANSWER: I am not sure, as he never
annotated the game, to the best of my
knowledge. He may have been worried
about ♗xg2 15.♔xg2 ♕d5+ , when the
knight cannot come to e4 and will
apparently be driven back to a lesser
square after ...h6.
However, here White can calmly offer the
queen exchange with 16.♕f3 ♕xf3+
( 16...h6? 17.♘xe6+ wins ) 17.♔xf3
when he still has a clear advantage – he
t h r e a t e n s t o t a k e o n b 5 , w h i c h B la c k
c a n n o t r e a l l y p r e v e n t , e x c e p t b y b4
, a f t e r w h i c h 18.cxb4 ♘c6 19.♖ac1
♘xd4+ 20.♔g2! just looks good for
White, ( but not 20.♔e4? ♘b3
, which allows Black tricks, based on a
knight fork at d2. )]
Bronstein's move avoids this queen exchange
variation and just prepares to push the f-pawn,
which will give a fierce attack.
14...c6? QUESTION: A strange move!
ANSWER: Yes, I am not sure what Korchnoi
was thinking here. After the text, which kills
his queen's bishop and knight at a single
stroke, White just sweeps his opponent away
on the kingside.
[ I think the only way to play is the
c o m p u t e r ' s 14...♗xg2 15.♘xg2 b4!
, seizing the chance to disrupt the white
p a w n c h a i n . N o w 16.cxb4 ♘c6
attacks two pawns, and Black seems to be
holding satisfactorily. ]
15.f4 ♕e7 16.♕d2 ♔g7 QUESTION:
Black seems to have established a blockade
on the light squares.
ANSWER: Yes, but Nimzowitsch wrote that
where an incorrect blockade is established,
the opponent will break it down precisely at
what appears to be its most fortified point.
 Bronstein − Move
79 by Move
That is what happens here – f5 looks
impenetrable, but...
17.f5! exf5 18.♖xf5! It is as simple as that.
The knight fork on f5 means that the rook is
immune, so White has succeeded in opening
the f-file and can double rooks next move.
His attack is crushing.
♗c8 The bishop prepares to transfer to e6, to
defend f7.
19.♖f6 ♖a7 20.♖af1 ♗e6 21.♕g5
Intensifying the pressure on the dark squares
around the black king.
♕e8
[ Note that 21...h6? loses at once.
EXERCISE: Can you find a win here?
ANSWER: White mates with 22.♖xg6+!
fxg6 ( or 22...♔h7 23.♖g7# ) 23.♕xg6# . ]
22.♗e4 The last white piece is included in the
attack.
♖g8 23.♘g2
[ This is good enough, but the computer
shows a more incisive line, which is also
quite instructive: 23.♖xe6! ♕xe6 24.♖f6
♕e8 .
QUESTION: But what has White gained?
He has just given up one of his attacking
rooks for the black bishop.
ANSWER: Yes, but he has removed the
bishop on e6, which was a key defender.
The important point to understand is one
which is similar to what we discussed in
the context of Bronstein-Ljubojevic, and is
again something Kasparov has emphasized
when demonstrating his games. All White
has done here is exchange off one
attacking piece for one defensive piece. In
addition, the attacking rook on f6 has been
replaced by its colleague, whereas there is
no replacement available for the black
bishop that was on e6, where it was doing
a vital job of defending f7 and f5. And,
indeed, White can now exploit its absence
Bronstein − Move by Move
to give mate in four, starting with
25.♖xg6+! a n d n o w e i t h e r hxg6 ( or
25...fxg6 26.♕f6+ ♔h6 27.♘f5+ ♔h5
28.♕h4# ) 26.♘f5+ ♔h8 27.♕h6# . ]
Bronstein's move, although less incisive, is
motivated by the same idea as 23 Rxe6!,
namely the desire to eliminate the defending
bishop. The white knight transfers to f4,
from where it threatens not only g6, but also
e6 and h5. The computer confirms that Black
has no defence.
23...♔h8 24.♘f4 ♖e7 25.axb5 axb5
26.♕h6 ♘d7 This just loses c6, but Black
has no moves.
[ QUESTION: What about challenging the
queen with 26...♕f8 ?
ANSWER: That loses at once to
27.♘xg6+ ( and even 27.♕xf8 ♖xf8
28.♘xe6 is good enough. )]
[ Q U E S T I O N : S o , w h a t a b o u t 26...♖g7
instead?
A N S W E R : T h e n a g a i n , 27.♘xg6+!
is decisive: fxg6 ( or 27...♖xg6 28.♗xg6
,etc ) 28.♖f8+ . ]
27.♗xc6 b4 Black could have resigned with
a clear conscience, but probably want to
extend the game beyond move 30.
28.d5 ♗g4 29.e6! ♕f8 30.♕xf8 ♘xf8
31.h3 Very sadistic.
[ Of course, many moves win, including the
crude captures 31.exf7 ♖g7 32.cxb4 . ]
31...g5
[ 31...♗f5 32.exf7 ♖g7 33.g4 is curtains. ]
32.d6
[ Again, most roads lead to Rome, such as
32.♖xf7 . ]
32...♖xe6 33.♘xe6 ♗xe6 34.cxb4 ♔g7
35.♗e8 c3 36.♗xf7 Certainly one of the
most one-sided hammerings Korchnoi can
ever have suffered.
1-0
 Bronstein − Move
80 by Move
29 E17
Bronstein,D
Keene,R
Teesside (Game 20) 1975
[Steve Giddins]
1.c4 Bill Hartston, who played in the same
tournament and later annotated the game in
the BCM, said that Bronstein apparently
chose this move to mark the fact that this
was his first game back in England in over
20 years (Hastings 1953/54 had been his last
visit). He even wrote it down in English
descriptive notation. Ray Keene himself has
confirmed both of these points to me.
♘f6 2.d4 QUESTION: Not the usual follow-
up, is it?
ANSWER: No, White just transposes back
into a regular 1 d4 opening. This move order
supports the theory that Bronstein's choice of
1 c4 was indeed just motivated by wanting to
mark his return to Blighty.
e6 3.g3 The Catalan System, nowadays very
popular at all levels. White aims to develop
pressure against the central light squares,
with his bishop raking down towards the
black queenside.
d5 4.♗g2 ♘bd7!? QUESTION: Is that
usual here?
ANSWER: Not especially, but it is a
perfectly good move.
[ Black has a key choice to make here. He
has two basic philosophical approaches to
the defence of t h e Catal an. One is t o
maintain his pawn on d5, so as to try to
shut out the white bishop on g2, and then
to develop his pieces within his own camp.
The other option is to capture on c4 with
4...dxc4 , clearing the long diagonal.
QUESTION: But doesn't that play into
White's hands, by opening the long
diagonal for his bishop?
Bronstein − Move by Move
ANSWER: Well, to some extent, perhaps,
but Black also frees his position. His idea
is to develop and get in the break ...c5
fairly quickly, trying to free his game. As
for the Catalan bishop, the argument is that,
if Black can evacuate the long diagonal
and develop his pieces satisfactorily around
it, the bishop will just be left looking
down an empty diagonal.
QUESTION: So which approach is
better?
ANSWER: There is no clear consensus and
it is largely a matter of taste. The open
lines are OK for Black, if he can develop
his queen's bishop, usually by getting in ...
b6 or ...b5 and then ...Bb7 If he cannot do
that, the danger is that his queenside will
never be developed satisfactorily, because
of the pressure against b7. ]
[ Keene's move presages a closed set-up,
whereby Black holds the d5 point.
However, his move is still a little unusual
in a move order context. Generally, the
closed system is introduced either by
4...♗e7 ]
[ or 4...♗b4+ 5.♗d2 and only now ♗e7
. The latter is the most popular nowadays,
the argument being that the white bishop on
d2 is slightly less well-placed than on c1,
as it no longer has the option of b3 and
Bb2. ]
5.♘d2 This is also slightly unusual, as White
does not generally commit his knight so early,
preferring to develop the kingside and hope
to be able to place the knight more actively
on c3.
QUESTION: So why did Bronstein play 5
Nd2 then?
ANSWER: I think it was a reaction to
Black's last move. There was a threat to take
on c4, and then defend the extra pawn by ...
Nb6, making it harder for White to regain it.
 Bronstein − Move
81 by Move
He can probably force recovery of the pawn,
but it will take a tempo or two, and will also
involve exchanging an extra pair of pieces,
which will relieve Black's slightly more
cramped position, so Bronstein decides
instead to avoid such issues.
♗e7 6.♘gf3 0-0 7.0-0 b6 Both sides are
developing logically. Having dealt with his
kingside, Black now brings the queen's
bishop to the long diagonal, neutralizing the
enemy Catalan bishop.
QUESTION: So has Black equalized?
ANSWER: Yes, I don't think he has any real
problems here.
8.b3 ♗b7 9.♗b2 c5 10.e3 QUESTION: It
all looks very cagey stuff by both sides!
ANSWER: Well, as we will see, White is
not devoid of ambition in this game. But for
t h e m o me n t , h e k e e p s h i s p o w d e r d r y ,
simply developing his pieces and
maintaining the tension. Such structures are
actually very rich, because the quartet of
central pawns allows many different
structures to arise. White could take on d5
and/or c5, and Black can do the same on d4/
c4.
QUESTION: I always find such positions
very confusing!
ANSWER: I think that is true of many club
players. The numerous possible pawn
exchanges mean that one must constantly
evaluate the consequences of the different
choices, and many weaker players tend to
prefer to clarify the situation, by making
early captures themselves. But amongst
grandmasters and masters, one will
frequ ently see the ten s i o n m a i n t a i n e d .
Bronstein understands that taking on d5
would bring him nothing, and instead, he is
prepared to recapture with pawns on both d4
and c4, accepting the so-called 'hanging
pawns'. We will see this later in the game.
Bronstein − Move by Move
♖c8 11.♖c1 ♖c7 12.♕e2 ♕a8
QUESTION: Wow, that looks pretty
extravagant!
ANSWER: Actually, it is a perfectly good
manoeuvre, despite the result of the game.
Reti showed back in the 1920s that in such
double-fianchetto set-ups, arising from Reti's
Opening and similar hypermodern systems,
the queen can often be effective when
backing up the bishop down the long
diagonal in this way. Black anticipates later
exchanges in the centre, which will open the
long light-square diagonal.
13.♖fe1 ♖fc8 This is the other point of
Black's set-up. His rooks double on the c-file,
preparing to give White the hanging central
pawns, and then target the c4-pawn.
QUESTION: So, is this good?
ANSWER: It is perfectly reasonable, but
with the white knight on d2, the pawn on c4
will be better defended than would be the
case were the knight on c3. However, one
can imagine the possibility of ...cxd4 and ...
dxc4, followed by ...Bb4, trying to remove
the knight and undermine the white c4-pawn.
QUESTION: That all sounds quite dangerous.
What is White going to do?
ANSWER: Well, the notable thing, if you
look at this position, is that Black's heavy
pieces have deserted his king – the queen
and both rooks are over on the queenside,
whilst White has various pieces, notably his
queen's bishop, eyeing up the enemy kingside.
That is the clue to what may happen. Black
can try to target the white pawns on the
queenside, but White will attempt to exploit
the lack of defenders on the kingside. After a
slow build-up, the stage is set for a possible
explosion.
14.cxd5 The first central pawn exchange.
Now Black must decide how to recapture.
♗xd5
 Bronstein − Move
82 by Move
[ 14...exd5 is the alternative, but that would
fix the d5-pawn in place and leave the
black queen and b7-bishop looking a
little silly, staring at the brick wall on d5.
Keene therefore prefers to keep the long
diagonal open.
QUESTION: So was that the right
decision?
ANSWER: Yes, I think it is fine. He would
be slightly worse after the pawn recapture
on d5, although it is not much, but taking
with a piece on d5 should be objectively
fine. ]
[ QUESTION: What about the knight
recapture?
ANSWER: That is also fine. After
14...♘xd5 15.e4 ♘5f6 16.e5 ♘d5
17.♘e4 , Black can equalize comfortably
with cxd4 , forcing exchanges on the c-
file. ]
15.e4 ♗b7 16.d5!? QUESTION: What is
this?
ANSWER: Bronstein sacrifices a pawn, to
open up the long diagonal for his dark-
squared bishop and also the central d- and e-
files for his pieces. It is a thematic type of
sacrifice in such positions, and here it is
motivated particularly by the fact that the
black queen is out of play on a8.
QUESTION: So are you saying White is
better?
ANSWER: Actually, no. Objectively,
chances are still balanced. White will get
compensation for the pawn, in the shape of
his active pieces, but Black is still fine.
exd5 17.exd5 ♗f8!? QUESTION: Why
didn't he take the pawn? Is he losing the
bishop on e7?
[ ANSWER: No. He could in fact have
t a k e n o n d 5 w i t h 17...♘xd5
. T h e n H a r t s t o n g i v e s 18.♘e5
as "unpleasant for Black", but after the
Bronstein − Move by Move
computer's
A) Hartston cites as an example of
Black's alleged difficulties the line
18...♘xe5? 19.♗xe5 ♖d7 , but now
20.♕g4 wins out of hand, ( as opposed to
H a r t s t o n ' s 20.♗h3 , which is also good
for White, but less convincing. );
B) 18...♗f6 Black is fine. ]
[ 17...♖e8 is another good option
(recommended by Hartston, in fact), when
Black is again fine. ]
18.♘e4 ♘xe4!? Now the attack starts to
develop with alarming speed.
QUESTION: What is wrong with capturing
on d5?
[ ANSWER: Well, with which piece? If we
s t a r t w i t h 18...♘xd5 we get another nice
illustration of the potential dangers in such
a position. White responds with 19.♘fg5!
and, suddenly, he has a very strong attack.
The threat is 20 Qh5, so Black really needs
to put one of his knights on f6 to stop this:
♘7f6 ( 19...♘5f6 20.♕c4!
is very unpleasant, when Black has no good
way to defend f7) 20.♘xf6+ ( the simple
20.♖cd1 may be even better) 20...♘xf6
21.♗xb7 ♕xb7 22.♗xf6 gxf6 23.♘e4
♕c6 24.♕g4+ ♔h8 25.♕f4 gives White a
strong initiative for the pawn. ]
[ QUESTION: So is the bishop capture
better?
ANSWER: Probably slightly better, yes,
but even here, after 18...♗xd5 19.♘fg5
Black has to be very careful. For example,
after a naive 19th move, such h6? ( instead,
19...♘xe4 20.♘xe4 ♖e8 is more sensible,
but even then, 21.♕g4 leaves Black facing
some nasty threats ) , there follows 20.♗xf6
A) 20...gxf6? 21.♘xf7! is killing: ♔xf7
( or 21...♗xf7 22.♘d6 ♗d5 23.♕d3
♗xg2 24.♕g6+ ♗g7 25.♖e7
and mates ) 22.♘d6+ ♗xd6 23.♕h5+
 Bronstein − Move
83 by Move
♔g7 24.♗xd5 and Black has to surrender
the queen to avoid a quick mate;
B) 20...hxg5 21.♗b2 , with a clear
advantage.
QUESTION: I don't understand how all
this has happened. One minute, we had a
quiet manoeuvring position, where you
were saying Black was fine, and now
suddenly he is being mated in variation
after variation. What is going on?
ANSWER: The thing is that, as I pointed
out earlier, Black's manoeuvring on the
queenside did leave his king's wing
somewhat bereft of defenders. Once
things open up, the white pieces suddenly
jump out and it only needs one error for
Black's king to come under a fierce attack.
That is what can happen in such
positions. ]
19.♕xe4 ♗xd5 20.♕f5!?
[ This works well enough, but the computer
gives 20.♕d3! as even stronger, when
White has more than enough compensation
for th e pawn, and Black is in serious
trouble. 21 Ng5 is the threat, and after, for
example, g6? ( 20...h6? does not stop the
threats: 21.♘g5! hxg5 22.♗xd5 ♕b8
23.♕g6! and Black has no way to defend f7,
as his knight does not have a single safe
square to move to) , there follows simply
21.♖cd1 and Black is actually completely
lost, as ♗c6 22.♕c3 f6 23.♖xd7 ♖xd7
24.♕xf6 ♗g7 25.♕e6+ ♔h8 26.♘e5!
is murderous. Once again, we see the
white pieces leaping out of their apparent
torpor and suddenly creating lethal
threats. ]
20...g6 Once again, the threat of 21 Ng5
leaves Black little choice.
21.♕f4 QUESTION: It does not look as
though Black should be in so much trouble
here.
Bronstein − Move by Move
[ ANSWER: Believe me, he is. The problem
is that now White threatens just 22 Rcd1,
followed by Ng5 and/or Rxd7 and Qe5,
with mate down the long diagonal. For
example, 21.♕f4 ♗g7 ( or 21...♕b8
22.♘g5! , which wins at once, because d5
is hanging,; and if Black just 'passes', with,
say, 21...a5 , then 22.♖cd1 ♗c6 23.♘g5
f6 24.♖xd7! ♖xd7 25.♕xf6 and it is all
over ) 22.♗xg7 ♔xg7 23.♖e7 and 24 Ne5
is a deadly threat. ]
21...f6 A horrible weakening, but, as the
previous note has shown, it is forced. He
simply must stop moves such as Ne5 and
Ng5.
22.♖cd1 Beautifully calm, bringing the last
white piece into the attack.
c4 The computer confirms that there is
nothing better. This move introduces the
threat of 23...c3, closing one of White's
deadly bishops, but now the move Black
thought his 21st move had prevented turns
out to be playable after all.
23.♘g5! ♖c5 Keene's move defends the
bishop on d5, but it is not enough to hold.
QUESTION: What is going on?
[ ANSWER: The main point is that
23...fxg5 now loses to 24.♕d4 , neatly
exploiting the drawback of Black's 22nd
move, which gave up control of d4: ♗xg2
25.♕h8+ ♔f7 26.♕xh7+ ♗g7 27.♕xg7#
.]
[ QUESTION: But what about taking on g2
instead?
ANSWER: After 23...♗xg2 there follows
24.♖xd7! (eliminating the defender of f6)
♖xd7 ( the computer shows as relatively
best 24...♗h3! 25.♘xh3 ♖xd7 26.♕xf6
c3 27.♗xc3 ♖xc3 28.♕xc3 and White has
an extra pawn, which should suffice to win,
although some technical problems remain)
25.♕xf6 when a lovely variation, pointed
 Bronstein − Move
84 by Move
out by the computer, is c3 ( 25...♖g7
26.♕e6+ is a smothered mate) 26.♕e6+
♔h8 27.♕xd7 ♕b7 28.♗xc3+ ♖xc3
29.♕d4+ ♕g7 30.♖e5! ♖c5 31.♔xg2
♖xe5 32.♕xe5 ♕xe5 33.♘f7+ ♔g7
34.♘xe5 and the upshot of all White's
brilliance is a minor piece ending with an
extra pawn, which he should win. ]
24.♖xd5!
[ 24.♕h4 is equally deadly (if fxg5 25.♕d4
), but Bronstein was naturally not going
to refuse the chance of a further sacrifice. ]
24...♖xd5 25.bxc4 QUESTION: Beautiful!
ANSWER: Yes, this calm pawn capture is
superbly impudent, just pinning the black
rook down the long diagonal. The white
bishops have certainly earned their corn in
this game.
fxg5 26.♕g4 Now he threatens simply mate
in one with 27 Qe6, as well as the win of the
queen by 27 Bxd5+. It is all over.
♖f5 27.♗xa8 ♖xa8 28.♕e4 ♖d8 29.♕d4
♔f7 30.g4!
[ A nice final touch. If 30.g4 ♖f4 ( or
30...♖c5 31.♕h8 and there is no longer the
defence ♘f6 ) 31.♕d5# .
A blistering attack, seemingly from a clear
sky, and a lovely illustration of the
dynamic potential contained in such
outwardly restrained set-ups. ]
1-0
30 E29
Bronstein,D
Najdorf,M
Budapest Candidates (Game 21) 1950
[Steve Giddins]
1.d4 ♘f6 2.c4 e6 3.♘c3 ♗b4 4.a3
The S‫ن‬misch variation, which although it
has never been regarded as the strongest
Bronstein − Move by Move
response to the Nimzo-Indian, is certainly
the most direct and challenging one.
QUESTION: Why is that?
ANSWER: Because it directly challenges
Black to implement his main idea of taking
on c3 and doubling the white pawns.
Effectively, White is announcing to his
opponent that he believes the latter's entire
strategic plan is faulty, and that the
possession of the bishop-pair, plus the
strengthened pawn centre, will more than
outweigh the weakness of the doubled pawns.
Readers of my companion volume
'Nimzowitsch: Move by Move' will be aware
that the great Nimzowitsch used frequently
to take on c3, even without the provocation
a2-a3, so great was his faith in the black
position. S‫ن‬misch players are the other
extreme, willing to spend a tempo forcing
the very exchange that Black intends anyway.
QUESTION: And what is the truth about the
variation?
ANSWER: It is probably not the best hope of
any advantage. With accurate play, Black
can neutralize the white threats and obtain a
good game. But it remains a very viable
weapon, and one which is very dangerous
indeed against an unprepared opponent.
[ At the time of writing, a variant of the
S‫ن‬misch is currently very popular at top
l e v e l . T h i s b e g i n s w i t h t h e m o v e 4.f3
, but in most cases, White follows up with
a rapid a3, and the two lines frequently
transpose. ]
4...♗xc3+ 5.bxc3 c5 QUESTION: Is this
best?
ANSWER: It is generally regarded as the
best move, yes. Black immediately fixes the
enemy doubled pawns and stakes a claim in
the centre. However, other moves are
possible.
[ Q U E S T I O N : W h a t a b o u t 5...d5
 Bronstein − Move
85 by Move
? That looks very natural.
ANSWER: It is quite thematic, in that it
continues the battle for the e4-square,
which is the cornerstone of Black's strategy
in the Nimzo-Indian, but the move has
been frowned upon ever since the famous
game Botvinnik-Capablanca, AVRO 1938.
The problem with ...d5 in these positions is
that White is able to undouble his pawns,
which Black usually wishes to prevent.
After 6.cxd5 exd5 7.e3 we transpose to
the Botvinnik-Capablanca game (which
actually arose via 4 e3 d5 5 a3, etc).
White has the bishop-pair and central
pawn majority, and will develop with Bd3,
Ne2, 0-0 and then play f3 and e4, trying to
get a pawn roller going.
QUESTION: So this is bad for Black?
ANSWER: Well, in fact, there are some
unanswered questions about the
Botvinnik-Capablanca game, and I am
not convinced that the position is as bad
for Black as is generally assumed, but that
is another story. Certainly, Black is
taking some risks playing this way, unless
he has a very specific idea in mind. ]
[ Instead, 5...b6 was a favourite of the great
Tigran Petrosian. Black invites his
opponent to establish a powerful pawn
c e n t r e a f t e r 6.f3 ♗a6 7.e4 ♘c6
a n d p u t s h i s o w n f a i th in r ap id p iec e
development. This is a very interesting
system, which is perhaps somewhat
underrated by theory. ]
6.e3
[ 6.f3 is another approach, and transposes
into the 4 f3 line referred to above, which
is currently popular. Instead, Bronstein
plans to develop with Bd3, Ne2 and 0-0,
before setting his central pawns in motion. ]
6...♘c6 7.♗d3 0-0 8.♘e2 d6!?
QUESTION: Why do you frown on this
Bronstein − Move by Move
move?
ANSWER: It is not bad, but Bronstein
himself noted at the time that there was no
immediate need for Black to move his d-
pawn yet.
9.e4 ♘e8 QUESTION: What is that about?
It looks a bit retrograde.
ANSWER: Actually, it is a standard idea in
the S‫ن‬misch, and was first practiced by
Capablanca, I believe. The point is to avoid
the pin by Bg5, which would be quite nasty
for Black – it would be hard to break, other
than by playing ...h6 and ...g5, which with
his king castled short, would be extremely
risky. There is also another point to ...Ne8,
although you would not know it from this
game. More on that in a moment.
10.0-0 b6 QUESTION: What is Black's idea
now?
ANSWER: He is pursuing another standard
plan in these positions. He intends to attack
the weak pawn on c4, by means of ...Ba6 and
...Na5.
11.f4 ♗a6?! QUESTION: Why the
punctuation? Isn't this standard?
ANSWER: It is the standard plan, yes, but I
think Black should attend to matters on the
kingside first.
[ I mentioned above that 9...Ne8 had
another motive, as well as avoiding the pin
b y B g 5 . T h a t i s t o p l a y 11...f5
in such positions, preventing the further
advance of the white f-pawn. This
prophylactic idea is another key one in
these structures. The course of the present
game shows what happens if Black fails to
take seriously the threat of f5-f6. In the
same event, the game Stahlberg-Reshevsky
continued (after 11...f5), 12.d5 ♘a5
13.dxe6 ♗xe6 14.exf5 ♗f7 ( 14...♗xf5
i s a l s o p o s s i b l e) 15.g4 ♘c7
, with sufficient compensation for the pawn.
 Bronstein − Move
86 by Move
The weakness on c4 will soon drop off. ]
12.f5! Immediately putting his finger on
Black's error. Now there is a very unpleasant
threat of the pawn pushing on to f6 in some
variations, as well as a possible exchange on
e6. Black is in trouble.
QUESTION: It seems odd that things should
have turned against him so suddenly, with
only a minor inaccuracy on his part.
ANSWER: To a considerable extent, that is
the nature of the S‫ن‬misch. In such a sharp
and doubled-edged variation, the cost of an
error is always much greater than in a quiet
position. White has some real trumps here,
in the shape of his bishop-pair and powerful
pawn centre. If he is allowed to exploit these,
then Black's position can become critical
very quickly. On the other hand, as several
o f t h e g a mes i n m y N i m z o w i t s c h b o o k
showed, the reverse is also true – White has
some potentially quite serious defects in his
position too, especially his doubled pawns,
and if he plays too slowly and allows his
opponent to block up the position, then
things can turn bad for him very quickly, as
well.
e5
[ 12...f6 13.fxe6 is impossible, so Black
tries to close the game. Now he hopes to
be able to play 13...f6, but Bronstein does
not give him time. ]
13.f6! ♔h8 QUESTION: Can't the pawn be
taken?
[ ANSWER: Not really. 13...gxf6? 14.♗h6
is obviously deadly, ]
[ whilst 13...♘xf6 14.♗g5 is also very
unpleasant for Black, in view of the threat
Ne2-g3-h5. Now taking on d4 leads to a
quick mate after cxd4 15.cxd4 ♘xd4?
( 15...♕c8 is more tenacious, but after
16.♗xf6 gxf6 17.♘g3 the position is
obviously dreadful for Black) 16.♘xd4
Bronstein − Move by Move
exd4 17.e5 dxe5 18.♗xf6 gxf6 19.♕g4+
♔h8 20.♕f5 . ]
With the text, Najdorf was hoping to create
counterplay on the open g-file, but he was
always an incorrigible optimist, and no such
play ever materializes. Instead, he is simply
battered on the kingside.
14.d5 Seizing the chance to relieve the
pressure against d4 and drive the knight
offside (on the previous move, 13 d5 would
have allowed the knight to swing over to the
kingside via e7).
♘a5 15.♘g3 QUESTION: That is a
surprise! Why not take on g7?
[ ANSWER: That would be equally good.
Bronstein suggests that he did not want to
encourage the black knight to g7, in a line
such as 15.fxg7+ ♘xg7 16.♘g3 f6
, but it still seems that White has a decisive
positional advantage after 17.♗h6 ♖g8
18.♕e2 . ]
15...gxf6
[ QUESTION: This looks like suicide!
Surely he had to keep things closed with
15...g6 , didn't he?
ANSWER: Yes, that was the only hope,
although objectively, after 16.♗h6 ♖g8
17.♕e2 , White's positional advantage is
overwhelming. ]
16.♘f5 ♗c8
[ 16...♖g8 is slightly more tenacious, but
t h e n s i m p l y 17.♘e3 kills Black's
counterplay against c4, and White can
quietly build up with Qe2, Rf3, Bd2, Raf1,
etc. ]
17.♕h5 Intending simply Rf3-h3.
♗xf5 18.exf5 ♖g8 19.♖f3 ♖g7
[ Bronstein points out a nice finish after
19...♘g7 .
EXERCISE: Can you find the win?
ANSWER: 20.♕xh7+ ♔xh7 21.♖h3+
♘h5 22.♖xh5+ ♔g7 23.♗h6+ ♔h7
 Bronstein − Move
24.♗f8# . ]
20.♗h6 ♖g8 QUESTION: Was there
anything remotely better?
[ ANSWER: Toran/Eliskases indicate that
20...♕e7 hangs on slightly longer, but that
then, instead of taking the exchange, White
continues 21.♖h3! ♕f8 22.♖g3 ♕g8
23.♗xg7+ ♘xg7 24.♕h6 and f6 drops.
Note that throughout this and other
variations, the knight on a5 is a mere
spectator, as is the black queen's rook. ]
21.♖h3
[ The threat is 22 Bf8, and after 21.♖h3
♘g7 22.♕h4 there is no defence to 23 Bg5.
A remarkable th rashin g o f o ne of th e
world's best players, and a very good
illustration of how dangerous the
S ‫ن‬m i s c h v a r i a t i o n c a n b e , a g a i n s t
inaccurate defence. ]
1-0
31
Bronstein,D
Szabo,L
Budapest Candidates (Game 22)
[Steve Giddins]
1.d4 ♘f6 2.c4 e6 3.♘c3 ♗b4 4.a3 ♗xc3+
5.bxc3 0-0?! QUESTION: Why don't you
like this move? It looks perfectly natural.
ANSWER: Maybe, but the early commitment
of the king creates a few problems for Black.
In particular, the pin with Bg5 becomes
rather more of a problem for him, as chasing
away the bishop with ...h6 and ...g5 is
obviously much less attractive now.
[ We saw the usual and best move 5...c5
in the previous game. ]
6.f3! Bronstein immediately seizes on
Black's inaccuracy, to set up a strong pawn
centre at once. After 5...c5, this plan of an
Bronstein − Move by Move
87 by Move
immediate 6 f3 is less effective and, indeed,
we saw Bronstein dispense with it in favour
of the slower build-up 6 e3 in the previous
game, but here, Black's having castled so
early helps White.
♘h5!? QUESTION: Hmm. That looks
strange. I thought one was not supposed to
move the same piece several times in the
opening? Why doesn't Black just develop his
other pieces?
ANSWER: The problem is that White is
threatening to set up a broad pawn centre
with e4, followed by Bg5 and aggressive
intentions on the kingside. Black really
needs to find a way to fight against this.
QUESTION: But how does decentralizing
the knight help? Surely it makes e2-e4 easier
to play, rather than more difficult?
ANSWER: Black's move has two points. The
first is to set up the incidental threat of ...
Qh4+. The second is to free his f-pawn to
advance to f5, trying to restrain White's
E27 central expansion.
[ The direct way to counter the central
expansion is 6...d5 , but as we discussed in
1950 the previous game, this allows White to
u n d o u b le h i s c- p a w n s a n d o b ta i n t h e
mobile central pawn majority, of
Botvinnik-Capablanca, after 7.cxd5 exd5
8.e3 , etc. ]
7.♘h3
[ 7.e4? ♕h4+ is awkward, since 8.g3?
runs into ♘xg3 . White is therefore forced
to bring his knight out to h3, so as to be
able to meet ...Qh4+ with Nf2. ]
7...f5 This is Black's second point. His f-
pawn takes part in the battle to prevent e2-e4,
which now seems impossible, because of the
check on h4.
8.e4! Nonetheless! Bronstein shows great
imagination and aggression in the early
stages of this game.
 Bronstein − Move
88 by Move
[ Interestingly, this same position after 7...
f5 was reached in 2014, in a game between
two of the top players of the current
generation, one of them the world
champion. The game S.Karjakin-M.Carlsen,
Gashimov Memorial, Shamkir 2014, saw
W h i t e s e t t l e f o r t h e m o d e s t 8.e3
, but after d6 9.♗e2 c5 10.0-0 ♘c6 11.g4
fxg4 12.fxg4 ♘f6 13.♘f2 h6 14.e4 e5
15.d5 ♘e7 Black had achieved exactly the
sort of blocked, Hübner-Nimzo-type
structure that he wanted. White eventually
hung on to draw, but he is already worse
here. ]
8...c5 QUESTION: But why didn't Black take
on e4? Surely that is the whole point of his
play?
[ ANSWER: Indeed, but presumably only
now did Szabo realize that things are not
so simple. In the tournament book, the
strong Hungarian GM Gideon Barcza gives
the line 8...fxe4 9.♗g5! ( of course not
9.fxe4? ♕h4+ , which is Black's idea)
9...♘f6 ( Barcza recommends as best
9...♕e8 10.fxe4 e5 , but here too, after
11.♗e2 ♘f6 12.0-0 White looks to be
clearly better, because ♘xe4? 13.♖xf8+
♕xf8 14.♕d3 is much too dangerous;
Black has his entire queenside asleep,
whilst all the white pieces are playing)
10.fxe4 h6 11.♗h4 g5 (this fails, but
otherwise, White has obtained his ideal set-
up) 12.♗d3!? ( in this line, Fritz points out
that the simple 12.♗g3 is probably better,
since ♘xe4 13.♕c2 d5 14.♘f2
is good for White) 12...gxh4 13.e5 ♘h7
14.♕h5 ♕e7 15.♔e2 ♕f7 ( if 15...♕g7
then 16.♖af1! , with the point that ♕xg2+?
17.♔e1 wins ) 16.♗g6 ♕g7 17.♖af1
with a strong attack. Sixty years on, the
computer basically agrees with this,
although it claims Black is holding after d6
Bronstein − Move by Move
.
QUESTION: So what is the conclusion
here?
ANSWER: The conclusion is that
Bronstein's 8 e4! is an excellent decision,
which fully accords with the demands of
the position. As we have emphasized in
the comments to the previous game, the
S‫ن‬misch is a committal system, in which
White accepts some major liabilities, in
order to obtain some dynamic trumps. He
needs to play energetically, in order to
make the most of his plusses. Bronstein's
handling of the position (almost certainly
improvised at the board) does exactly that,
whereas 60 years later, one of the world's
leading GMs was not up to the task and
settled for the tame 8 e3. This is certainly
food for thought for those who like to
claim that the modern generation of top
players are stronger than their predecessors
of past generations. ]
9.e5 QUESTION: How do you assess this
position?
ANSWER: I think White is already better.
His pawn centre is very powerful, the black
knight on h5 is offside, and the dark squares
in Black's camp are weak.
QUESTION: But White still has his doubled
c-pawns.
ANSWER: Yes, but they are not under any
pressure, are they? A weakness is only a
weakness if it can be attacked. Here, White
can easily cover c4 (e.g. by Bd3 and Qe2),
and meanwhile, the extra centre pawn
controls key squares such as d4 and d5.
♘c6 10.f4 g6 QUESTION: Why this
move?
ANSWER: It is not a move that Black
wanted to play, since it further weakens his
dark squares, but there was a threat of Be2,
and the knight would have had no retreat.
 Bronstein − Move
89 by Move
11.♗e2 b6 12.0-0!
[ N o t f a l l i n g f o r 12.♗xh5? ♕h4+!
. Now there really is a threat to take on h5,
however, so the knight must retreat. ]
12...♘g7 13.♗e3 QUESTION: Black's
position is looking pretty ropey!
ANSWER: It is absolutely revolting. White
has a crushing positional advantage, in fact.
His centre is enormously powerful, the
enemy dark squares are very weak, and
White has much more space. The rest is in
the nature of an execution.
cxd4 Undoubling the white pawns, but there
was no other way to save the pawn on c5.
QUESTION: Could he have sacrificed it, in
the hope of exposing the doubled c-pawn,
for compensation?
ANSWER: That is a good idea in principle,
but the problem is that, after dxc5, White
exposes the dark squares, such as d6, much
more, and also the d-file, including the
backward pawn on d7.
14.cxd4 ♗a6 15.♕a4 ♕c8 16.♖fc1
White's position is an absolute picture. He
has in mind something like Bf3 and
eventually d5, after due preparation.
♖b8 17.♖ab1 ♘a5 18.♘g5 ♕c6 19.♕b4
♕c7 20.d5 The centre finally advances. Note
how the black queenside pieces, especially
the knight on a5, are hopelessly offside –
just as in the previous game.
♘b7 EXERCISE: Now spot a good move
for White!
21.♕e1?!
[ ANSWER: The computer and, to his credit,
the silicon-free Barcza point out that 21.c5!
is more incisive, exploiting the loose
bishop on a6 and the potential pin on the c-
file. -- ( Now 21...♗xe2 loses material to
the zwischenzug 22.cxb6! ♕d8 23.bxa7
♖a8 24.♕xb7 and Black can resign,; whilst
21...bxc5 22.♗xa6 wins a piece,; as does
Bronstein − Move by Move
21...♘xc5 22.♗xa6 ♘xa6 23.♖xc7 ♘xb4
24.axb4 . )]
However, Bronstein's move does not actually
spoil anything, and retains White's crushing
positional grip on the game.
21...♘c5 22.♕h4 QUESTION: Switching
targets?
ANSWER: Yes, the classic two-weakness
strategy. Note that the reason this works is
because Black's position is effectively cut in
two by the pawn on d7, which prevents the
pieces on the queenside from influencing
events on the other wing.
h5 23.♗f3 Disposing of the threat of 23...
exd5, exploiting the pin along the a6-f1
diagonal.
♖bc8
[ 23...d6 is the last chance to prevent
White's next, but then 24.dxe6 dxe5
25.♗d5 ♕e7 26.♘f3 is a simple solution. ]
24.d6 Finally severing the black position
permanently in two.
♕d8 25.♕g3! QUESTION: What is the
idea?
ANSWER: Bronstein is just turning his
attention to the latest weakness to appear in
Black's position. By forcing 22...h5, White
created a major weakness on g6, and now he
takes aim at it.
♘d3
[ 25...♕e8 would defend g6 for the moment,
b u t t h e n s i m p l y 26.a4! is one rather
sadistic way of winning. The a-pawn is
immune ( ♘xa4? 27.♖a1 ) and meanwhile,
the pawn comes on to a5, undermining
the queenside. ]
26.♖c3 ♗xc4 27.♗b7 b5
[ T h e e x c h a n g e i s l o s t , s i n c e 27...♖b8
28.♖xc4 ♖xb7 29.♗d4! neatly traps the
errant knight on d3, thanks to the pin on the
b-file. ]
28.♗xc8 ♕xc8 29.♘f3 Having bagged some
 Bronstein − Move
90 by Move
material on the queenside, Bronstein calmly
returns to his original idea of picking off the
remains on g6.
♔h7 30.♘h4 ♕e8 31.♗d4 The knight on
d3 is again trapped, and Black finally called
it a day – doubtless a blessed relief, after
what must have been one of the worst
afternoons of Szabo's life. Many readers of
this book may know little or nothing of
Laszlo Szabo, but if I tell you that he played
in both the 1950 and 1953 Candidates'
tournaments, I hope you will get the
message: he was one of the best players in
the world in the decade or so after the end of
the Second World War. He simply cannot
have been crushed in such style very often in
his life.
1-0
32 E45
Botvinnik,M
Bronstein,D
17th matchgame, Moscow WCh. (Game 23)
[Steve Giddins]
1.d4 ♘f6 2.c4 e6 3.♘c3 ♗b4 4.e3 b6
5.♘ge2 ♗a6 This move, nowadays quite
standard, was actually a novelty in this game.
Bron stein claims he thought it up over
breakfast on the day of the game – a striking
contrast to modern world championship
matches, where such novelties are liable to
have been subjected to deep analysis months
in advance.
[ In the 13th matchgame, he had played
5...♗b7 , but failed to equalize. ]
6.a3
[ 6.♘g3 is also a popular alternative, and is
nowadays regarded as the best try for an
advantage. ]
6...♗e7
Bronstein − Move by Move
[ Following the improvement outlined in the
next note, the line 6...♗xc3+ 7.♘xc3 d5
became more popular here, when Black
cedes the bishop-pair, but can hope to
develop pressure against White's c- and d-
pawns. ]
7.♘g3?! QUESTION: This looks natural.
Why is it dubious?
ANSWER: It fails to offer White any
initiative, as the further course of the game
shows.
[ Three years later, in his world
championship match against Smyslov,
Botvinnik unveiled the improvement:
7.♘f4 d5 8.cxd5 ♗xf1 9.♔xf1 exd5
(all of which had been played in a game R.
Fine-G.Shainswit, New York 1951), and
now the flank blow 10.g4! , which sufficed
to put the entire line starting with 6...Be7
under a cloud for decades. ]
7...d5 8.cxd5 ♗xf1 9.♘xf1 exd5 10.♘g3
♕d7! QUESTION: So what is happening
here?
ANSWER: Black has emerged from the
opening with a good game. He has solved all
his development problems, and the exchange
of light-squared bishops has left White with
some weaknesses on the light squares in his
camp, as well as a rather poor bishop.
QUESTION: So why is Black's last move so
good?
ANSWER: Bronstein was very proud of his
play in this game, and gave a long note to
the last move, explaining how it was "the
start of a series of strong moves, which have
the aim of confusing the opponent and of
disorganizing his harmonious train of
thought".
QUESTION: What exactly does that mean?
ANSWER: I am not sure. The immediate
point of 10...Qd7 is to stop the white knight
coming to f5. In the more medium term,
 Bronstein − Move
91 by Move
White would like to mobilize his central
pawn majority with f3 and e4 – remember
Botvinnik-Capablanca! Bronstein plays
some subtle moves to meet this threat.
11.♕f3?! This is not terribly good, as
Botvinnik later acknowledged. He hoped to
exploit the f5-square, but this is not possible,
and meanwhile, the queen is misplaced for
White's intended plan of f3 and e4.
♘c6 QUESTION: Why does he block his c-
pawn like this? Isn't that supposed to be bad
in queen's pawn openings?
ANSWER: It often is, yes, but here Black has
a subtle purpose. He is anticipating White's
subsequent f3 and e4, and intends to stop
this by pressuring the d4-pawn. 11...Nc6 is
the start of that process.
12.0-0 QUESTION: So why doesn't White
come into f5 with his knight?
[ ANSWER: It does nothing, after simply
12.♘f5 ♗f8 followed by ...g6. ]
12...g6 13.♗d2 0-0 14.♘ce2 QUESTION:
White's play is very unimpressive!
ANS WER: Yes, Bo tvi nni k was clearl y
uncomfortable and is drifting.
[ S v e s h n i k o v r e c o m m e n d e d 14.b4
, either here or on the previous move. ]
14...h5! Seizing the initiative, and threatening
to drive back the knight from g3, thus further
undermining White's hopes of achieving e4.
It is interesting to compare Black's play in
this game with the opening game of the
Petrosian-Botvinnik match in 1963. In a
rather similar structure, Botvinnik, this time
as Black, implemented a very similar plan to
that used by Bronst ein here. Botvinnik
always learned from his defeats.
15.♖fc1 White frees the f1-square for his
knight, and also starts to create some
pressure on the c-file. The problem is that he
has lost a couple of tempi with his last few
moves, and the counterplay never
Bronstein − Move by Move
materializes.
h4 16.♘f1 ♘e4 17.♘f4
[ Else 17...Ng5 is a strong threat: for
example, 17.b4 ♘g5 18.♕f4 h3
and White's kingside is seriously
weakened. ]
17...a5! Another excellent move. Bronstein
tightens his grip on the entire board, by
preventing White's intended b4.
18.♖c2 ♗d8! QUESTION: What is that for?
It looks a little retrograde.
ANSWER: It is prophylaxis against White's
plan to pressure the c-file. Black defends c7
in advance, so his knight can move from c6.
It must be said that, if one were to see this
game without knowing which player was
which colour, one would certainly tend to
assume that Botvinnik was playing Black.
Bronstein plays with just the sort of
remorseless iron logic associated with his
opponent.
[ Bronstein himself comments in 'David
Bronstein: Chess Improviser' that White's
knight manoeuvre to f4 intended a rook
sacrifice on c7. What he presumably means
is 18...♘d8 19.♖xc7! ♕xc7 20.♘xd5
when White will regain a piece on either e7
or e4, with two pawns for the exchange. It
is notable that this is the computer's first
choice for White after 18...Nd8.
Bronstein's subtle bishop move sidesteps
it. ]
19.♗e1 ♘e7 20.♕e2 ♘d6 QUESTION:
What is the assessment? Is Black actually
better?
ANSWER: I would say he is, yes, although
the computer gives the position as
completely equal. Black certainly has the
initiative, and his manoeuvring is much
more purposeful. His knight on d6 is
excellently placed in this Carlsbad structure,
eyeing up e4 and c4.
 Bronstein − Move
92 by Move
21.f3 QUESTION: Is this good? It weakens
e3, after all.
ANSWER: Indeed it does, but therein lies
the dilemma. The advance e4 is White's only
real hope of serious counterplay, so he feels
obliged to attempt it, but there is always the
risk that, if he fails to get in e4 in
favourable circumstances, then f3 will just
prove to be a weakening, as in fact happens.
g5 QUESTION: Starting the attack?
ANSWER: Not exactly. Bronstein says it is
too early to speak of a direct attack, and that
Black still needs to bring up his forces. But
the move starts the process, by setting up an
eventual pawn offensive on the kingside, as
well as just defending the h4-pawn, which
White's last move attacked.
22.♘d3 ♕e6 Black consistently stops White
playing e4.
23.a4?! QUESTION: What does that do?
ANSWER: I assume Botvinnik was disturbed
by the threat of ...a4, followed by ..Nc4, but
this is not a very nice move to play. His
position is passive and unpleasant though,
and it is always hard to suggest constructive
moves in such positions.
[ Maybe 23.♖d1 makes more sense,
defending d4 and hoping to arrange Nf2
and e4, but Black will not sit idly by in
the meantime, of course. ]
23...♘g6 24.h3 f5 25.♗c3 Bronstein writes
that Botvinnik clearly feared ...Rf7, ...Bf6
and ...f4, so he overprotects d4. However,
the problem with this move is that he
relieves Black of any worries over his c7-
pawn, which permits the black queen's rook
to come to the e-file.
[ S v e s h n i k o v ' s s u g g e s t i o n o f 25.♖ac1
would at least make that plan more difficult
to carry out. ]
25...♗f6 26.♖e1 ♖ae8 27.♕d1 ♖f7 28.b3
♖fe7 29.♗b2 Black's build-up has been
Bronstein − Move by Move
completed and his pressure down the e-file
reaches its apotheosis. It is time for the
breakthrough. I am astounded to see that
Fritz continues to assess this position as dead
equal.
f4 30.♘e5?!
[ QUESTION: But the black queen is
attacked after 30.exf4 ?
ANSWER: Yup! Bronstein had a queen
s a c r i f i c e i n m i n d . H e i n t e n d e d ♕xe1
31.♘xe1 ♖xe1 32.♕d3 ( 32.♕d2 ♘xf4
is no better) 32...♘xf4 33.♕a6 ♘f5
with a winning position.
QUESTION: So what should White have
done? ]
[ ANSWER: Bronstein himself gives
30.♘f2 as "logical", but assesses the
p o s i t i o n i n B l a c k ' s f a v o u r a f t e r fxe3
31.♘g4 ♗g7 32.♖xe3 ♕f7 , because of
the weak d4-pawn and the active black
knights, which threaten to jump into f5 and
f4.
QUESTION: Do you agree?
ANSWER: Yes, although the computer
claims White is just equal here too. I
don't think this is so, but it is clear that
this is in any case what White should
have tried. Instead, Botvinnik jettisons a
pawn in desperation. It was noteworthy in
this match that whenever the position
became sharp and concrete, Botvinnik
tended to start going astray, partly because
he was usually in time-trouble by then. ]
30...♗xe5! QUESTION: That is not the way I
would have expected Black to recapture! I
would have taken with the knight.
[ ANSWER: In which case, you would have
fallen into Botvinnik's trap. After
30...♘xe5? 31.dxe5 ♗xe5 32.exf4
White is winning, because of the pin down
the e-file: for example, ♗d4+ 33.♕xd4
♕xe1 34.♕h8+ ♔f7 35.♕f6+ ♔g8 36.f5
 Bronstein − Move
93 by Move
and Black is busted. ]
31.dxe5 ♘f7 32.exf4
[ Bronstein gives this as a mistake, and says
32.♘h2 "was sharper". However, this just
looks bad after ♘gxe5 33.exf4 ♘xf3+
34.♘xf3 ♕xe1+ 35.♕xe1 ♖xe1+
36.♘xe1 ♖xe1+ 37.♔f2 ♖b1 and Black is
a pawn up in the ending. ]
32...♘xf4 33.♘h2 c5 34.♘g4 d4?
A time-trouble mistake that could have cost
him dearly.
[ Instead, 34...♕g6 keeps a large advantage,
the difference from the next note being that
now 35.♗c1 is met simply by ♘xe5 . ]
35.♘f6+?? A shocker of a blunder, which
ends a most interesting game.
[ Fritz still thinks Black is only slightly
b e t t e r a f t e r 35.♗c1 when, amazingly, it
suggests as best the striking move ♕d6
, pr ep ar in g to lif t th e e5 -p awn. After
36.♗xf4 gxf4 37.exd6 ♖xe1+ 38.♕xe1
♖xe1+ 39.♔f2 ♖e6 40.♖e2 ♖xe2+
41.♔xe2 ♘xd6 White does indeed have
very good drawing chances. ]
35...♕xf6 It is a pity that the mutual errors at
moves 34 and 35 spoil the game somewhat,
but I still thought it worth including, as
Bronstein's positional build-up earlier in the
game was very impressive. One does not
often see Botvinnik outplayed with the white
pieces with such iron logic.
0-1
33 E47
Bronstein,D
Keres,P
Gothenburg Interzonal (Game 24) 1955
[Steve Giddins]
1.d4 ♘f6 2.c4 e6 3.♘c3 ♗b4 4.e3 c5
5.♗d3 b6?! QUESTION: What is wrong
Bronstein − Move by Move
with this?
ANSWER: It is a slightly inaccurate move
order, which allows White to get in Bd3 and
Nge2. Usually, if Black is going to play ...b6,
he will do it at move 4, so as to stop White
getting that set-up.
6.♘ge2 ♗b7 7.0-0 cxd4 8.exd4 0-0
Allowing a strong central thrust, after which
Black has some problems.
[ QUESTION: So should he play 8...d5
instead?
ANSWER: Unfortunately, he can't, because
it loses a piece after 9.♕a4+ ♘c6 10.cxd5
♘xd5 11.♗b5 ♖c8 12.♘xd5 ♕xd5
13.♕xb4 . The text-move prepares 9...d5,
but Bronstein pounces first. ]
9.d5! h6 QUESTION: But isn't that pawn
hanging?
[ ANSWER: Only apparently. After 9...exd5
10.cxd5 ♘xd5 ( 10...♗xd5?? 11.♘xd5
♘xd5 12.♗e4 is even worse) 11.♘xd5
♗xd5 White regains the pawn with
12.♗xh7+ ♔xh7 13.♕xd5 with a clear
positional advantage.
QUESTION: Why is White better here?
ANSWER: Black has a weakness on d7 and
his kingside has also been weakened.
QUESTION: So the move in the game now
threatens to take on d5, right?
ANSWER: Exactly, because White would
no longer have the same Bxh7+ tactic to
regain the pawn. Bronstein therefore
defends d5 with his next move. ]
10.♗c2 QUESTION: So you think White is
better here?
ANSWER: Yes, slightly. His pawn centre c4-
d5 shuts out the bishop on b7 and prevents
Black staking a central claim of his own with
...d7-d5. Fritz assesses the position as
completely equal, but I do not believe that.
♘a6 Since Black has problems playing ...d6
at any stage, to bring the queen's knight to
 Bronstein − Move
94 by Move
d7 (his dark-squared bishop would be in
danger after ...d6, for one thing), this is the
only plausible way to develop the knight, but
it is clear that the black pieces are tripping
over each other to some extent. The knight
will not be able to come to c5, for example,
until the bishop on b4 has moved away (else
the bishop will be in danger of being trapped),
but it is also not clear where the bishop
should go, as putting it on e7 is impossible,
in view of d5-d6.
[ QUESTION: Can't he just play something
like 10...♗xc3 11.♘xc3 exd5 12.cxd5 d6
and then ...Nbd7?
ANSWER: He could, but with more space
and two bishops, White would just be better
here. ]
11.♘b5! An unpleasant move for Black to
face. The threat is 12 a3, trapping his bishop.
QUESTION: But that threat is easily met,
isn't it?
ANSWER: Yes, but Bronstein explains that
the move has a deeper purpose – to force
Black into exchanging on d5, when the f5-
square will become available for the white
knight.
exd5
[ 11...♖e8 is a way to avoid the exchange
on d5, but Keres evidently felt he should
challenge the correctness of White's
whole concept, by accepting the pawn
sacrifice. As will be seen, it all depends
on the evaluation of the position after
White's sacrifice at move 14. ]
12.a3 ♗e7 13.♘g3 Equine occupation of f5
is at the heart of White's idea.
dxc4?
[ Bronstein points out that taking the
second pawn was not compulsory, but after
13...♖e8 14.♕f3 ♘c5 15.♘f5 ♗f8
16.♕g3 White would still have a strong
attack, and at the cost of only one pawn. ]
Bronstein − Move by Move
[ However, Kasparov suggests as the most
logical continuation 13...♘c5 . 14.--
( Now he shows that the aggressive tries
14.♗xh6 gxh6 15.♕d2 ♘fe4! 16.♕xh6
♗f6; and 14.♘f5?! dxc4 15.♗xh6?! gxh6
16.♕d2 ♗e4! 17.♕xh6 ♗xf5 18.♗xf5
♘e6 are both inadequate for White,;
so he should instead settle for 14.cxd5
♗xd5 15.♖e1 "with lasting compensation
for the pawn" (Kasparov). )]
14.♗xh6!! QUESTION: Wow! Talk about
lightning from a clear sky!
ANSWER: Well, to a player like Bronstein,
there are a few clouds over the black king's
position. He is behind in development and
his knight on a6 is out of the game, whilst
White has several minor pieces pointing at
th e b lack king . But ev en so , this p iece
sacrifice is a big gamble, as White cannot
have foreseen any forced win or recovery of
ma t e r i a l . H e j u s t j u d g e d t h e a t t a c k i n g
chances as being worth the material
investment.
gxh6 15.♕d2 ♘h7? This position has been
the subject of deep analysis by numerous
commentators over the years, and the text
has generally been regarded as the decisive
mistake.
[ Instead, attention has focused on 15...♘c5!
as the best defence.
QUESTION: Why is that the move people
want to play?
ANSWER: It gets the offside knight into
the game. Now the knight can come to e4
or e6, defending the key square g7 in
various lines, or to d3, shutting out the
important white bishop on c2.
QUESTION: So what does White play?
A N S W E R : H e c o n t i n u e s 16.♖ae1
when Black should play ♘d3 , with a
critical position.
A) Vainstein records that Bronstein's first
 Bronstein − Move
95 by Move
idea was 17.♕xh6 ♖e8? 18.♖e5?
( instead, Kasparov points out that
18.♗xd3 cxd3 19.♘f5 is winning for
White, as the computer confirms: ♗f8
20.♕g5+ ♔h7 21.♘bd6 ♖xe1 22.♖xe1
and the rook comes to the third rank
w i t h d e c i s i v e e f f e c t) 18...♘xe5
( b u t t h e n h e r e a l i z e d t h a t 18...♘g4!
simply wins for Black) 19.♘h5 ♘xh5
20.♕h7+ ♔f8 21.♕h8# .;
B) QUESTION: So does that mean
17.♕xh6 wins for White?
ANSWER: Actually, no. Black has the
stronger defence ♘e8! , as pointed out
by the computer and Kasparov (the latter
presumably with the help of the former).
The analysis then continues 18.♖e5 f5
19.♕g6+ ♘g7 20.♘xf5 ♖xf5 21.♖xf5
♕e8 22.♕g4 d5 with an unclear
position.;
C) However, Bronstein had actually
decided over the board on a stronger
continuation of the attack, starting with
the exchange on d3: 17.♗xd3! cxd3 .
C1) N o w h e i n t e n d e d 18.♕xh6?!
, b u t t h i s i s a g a i n m e t b y ♘e8!
with unclear play: 19.♖e5 f5 20.♘xf5
♖xf5 21.♖xf5 ♘g7 , etc.;
C2) Stronger is 18.♘f5 as given by
Kasparov.
QUESTION: Delaying the capture on
h6?
ANSWER: Indeed. To paraphrase one
of Tony Blair's most notorious lines,
"This is no time for platitudes, but I feel
the threat is stronger than the execution.
" Now there is a maze of complicated
lines, including some lovely mates.
C2a) Thus, 18...♖e8? runs into
19.♘xh6+ ♔f8 20.♕g5
(simply intending 21 Nf5 and mate)
♗c5 21.♘f5 and now ♖e6 ( 21...♘h5
Bronstein − Move by Move
is the only way to avoid immediate
disaster, but just 22.♖xe8+ ♕xe8
23.♕xh5 is obviously much better for
White ) 22.♕g7+ ♔e8 23.♕h8+ ♗f8
24.♘bd6# is a delightful example of
what problemists call a double pin-
mate.;
C2b) QUESTION: So can Black do
better after 18 Nf5 then?
A N S W E R : Y e s . 18...♗e4
is the best defence. Now the next
critical position arises after 19.♘bd4
♖e8 20.♘xh6+ ♔f8 21.♕g5
. Kasparov devotes half a page of
solid analysis to this position, and
then adds that he does not pretend to
have exhausted the possibilities. But
his analysis suggests that the best
move for Black is ♗c5! , after which
he was unable to find a win. There is
little sense in reproducing all of his
analysis, but here is one typical line:
22.♖xe4 ♖xe4 23.♘df5 ♔e8
24.♕g7 ♕c7 25.♕xf6 d5 26.♕h8+
♗f8 27.♘g8 ♔d7 28.♘f6+ ♔e6
29.♘xe4 dxe4 30.♘g7+ ♔d7
31.♕h3+ ♔d8 32.♕f5 ♖c8 33.♘e6+
fxe6 34.♕xf8+ ♔d7 35.♕g7+ ♔d8
36.♕g8+ with a draw by perpetual
check.
QUESTION: So, the upshot seems to
be that Bronstein's sacrifice was only
good for a draw?
ANSWER: That seems to be the
conclusion, if all the analysis is
correct. But the position is so
complicated that even with a modern
engine, it is hard to be really
definitive. From a practical viewpoint,
the sacrifice gives enormous chances.
Keres was one of the best players ever,
and a noted defender in his mature
 Bronstein − Move
96 by Move
years, but even he was unable to cope
with the defensive task over the
board.;
C2c) 18...--;
C3) 18.--;
D) 17.-- ]
16.♕xh6 f5
[ Clearly forced, since 16...♘f6? 17.♘h5
mates at once. ]
17.♘xf5 ♖xf5
[ Again forced, since 17...♖f7 18.♕g6+
♔f8 19.♘bd6 is murder. ]
18.♗xf5 ♘f8 QUESTION: So White has
regained some material. How do you assess
the position?
ANSWER: White is basically winning. As
Kasparov sums it up: "White has almost
restored material equality, has totally
eliminated the black king's pawn cover and
retains a fierce attack, which will become
even more obvious after the rook from a1
joins in the assault".
19.♖ad1 ♗g5 20.♕h5 ♕f6 The black
queen comes to her king's aid and the
immediate threats are repulsed, but Black's
game remains hopeless.
21.♘d6 ♗c6 22.♕g4 Threatening 23 h3 or
23 f4.
♔h8 23.♗e4!
[ Kasparov hails this as the "key move",
a l t h o u g h F r i t z t h i n k s s i m p l y 23.♖fe1
is even better. But Bronstein's move is
more logical, as it creates two threats: 24
Qxg5 and 24 Bxc6, followed by 25 Ne4. ]
[ QUESTION: Hang on, though – can't
W h i t e j u s t w i n a p i e c e w i t h 23.♕xg5
at once?
ANSWER: This is less good, because after
♕xg5 24.♘f7+ ♔g7 25.♘xg5 ♔f6
Black regains the piece. White is still
much better, but Bronstein's choice is
more decisive. ]
Bronstein − Move by Move
23...♗h6
[ K a s p a r o v s u g g e s t s t h a t 23...♗f4
was possibly the best chance, but White
still seems to be winning comfortably after
24.♗xc6 dxc6 25.♘f5 ♕g5 26.♕h3+
♔g8 27.♖fe1 . Note how Black's queen's
rook has still not moved in this game,
whilst his queen's knight has moved only
once, to the offside square a6. ]
24.♗xc6 dxc6 25.♕xc4 The simplest. With
rook and two pawns against two minor pieces,
plus the naked black king, the rest is just a
mopping-up exercise for White.
♘c5 26.b4 ♘ce6 27.♕xc6 ♖b8 28.♘e4
♕g6 29.♖d6 ♗g7 30.f4 Threatening 31 f5,
winning more material.
QUESTION: Why doesn't Black just resign?
ANSWER: Who knows? Bronstein might
well have been in time-trouble here, and
besides, resignation never brings even a
fraction of a point. Where there is life, there
is hope.
♕g4 31.h3 ♕e2 32.♘g3 ♕e3+ 33.♔h2
♘d4 QUESTION: Why doesn't he take on
f4?
[ ANSWER: Good question. Vainstein
gives a complicated and not entirely
c o r r e c t v a r i a t i o n s t a r t i n g 33...♘xf4
34.♖h6+ , ( but a much simpler solution is
34.♕c7! ♖e8 35.♕xg7+! ♔xg7 36.♘f5+
, a variation that Bronstein himself
pointed out, back in the days when Fritz
was not even a glint in its programmer's
eye. )]
34.♕d5 ♖e8 35.♘h5
[ Hereabouts, there is some evidence of
time-trouble. Simpler was 35.♕h5+ ♔g8
36.♖xd4 ♗xd4 37.♘f5 and wins
(Kasparov). ]
35...♘e2 36.♘xg7?!
[ And here, 36.♕g5! ♕e7 37.♕g4!
wins at once. ]
 Bronstein − Move
36...♕g3+ 37.♔h1 ♘xf4
[ 37...♔xg7 is slightly more tenacious,
although White has many ways to win: for
instance, 38.♕b7+ ♔g8 39.♖d5!
with the horrible threat of Rg5+, and if
♘xf4 40.♖f5 ♖e7 41.♖xf8+ ♔xf8
42.♕b8+ and 43 Qxf4. ]
38.♕f3
[ QUESTION: Couldn't he have played
38.♘xe8 ?
ANSWER: Yes, that also wins. ]
38...♘e2 39.♖h6+
[ If 39.♖h6+ ♔xg7 40.♕f6+ ♔g8
41.♖h8# .
A wonderful game by Bronstein, which
won the best game prize at this Interzonal. ]
1-0
34
Bronstein,D
Golombek,H
Alekhine Memorial, Moscow (Game 25)
[Steve Giddins]
1.d4 ♘f6 2.c4 e6 3.♘c3 ♗b4 4.♘f3
QUESTION: After seeing the devastation
B r o n s t e i n i n f l i c t e d w i t h t h e S ‫ن‬m i s c h
variation in Games 21 and 22, this comes as
something of a surprise!
ANSWER: Well, I can see what you mean,
but tempus fugit, as they say down my local,
and theory advances. In the Candidates
tournaments of 1953 and 1956, Geller played
the S‫ن‬misch as White against Smyslov, and
was beaten in model style both times. Black
players had started to find the right set-up
against the variation, and Bronstein pretty
much stopped playing it after that. Besides, 4
Nf3 is a natural move, and was also less well-
studied at that time. This would have suited
Bronstein, playing against a weaker
Bronstein − Move by Move
97 by Move
opponent.
b6
[ 4...c5 is also a good move, but the text
transposes into a Queen's Indian, which
was another favourite defence with
Golombek. ]
5.e3 With this, the transpositional game
takes another turn, and we revert to a Nimzo-
Indian proper.
[ 5.♗g5 is more aggressive, and might have
been expected from Bronstein, but, as we
will see, he preferred to rely on
outplaying his opponent positionally. ]
5...♗b7 6.♗d3 ♘e4 QUESTION: Black is
moving a piece twice in the opening!
ANSWER: Yes, that is true, but despite this,
6...Ne4 has always been pretty much the
main line here.
[ However, 6...0-0 7.0-0 and now all of c5
E43 , ( 7...d5; and even 7...♗xc3 8.bxc3 ♗e4
are very respectable choices also. )]
7.0-0! QUESTION: Does this lose a pawn?
ANSWER: It sacrifices one, certainly, but
this is definitely the strongest move here.
Black, as noted above, has spent a tempo
moving his knight a second time. His aim is
to exchange on c3, doubling the white pawns,
and then to establish a grip on e4. If he can
do this and castle to safety, he will stand
well, and be in a position to set up a typical
Nimzo blockade with ...c5, etc. White
therefore needs to play energetically to
prevent this and to exploit his lead in
development. His seventh move does just this,
developing further and inviting Black to
waste more time pawn-grabbing.
[ Instead, the careful 7.♕c2?! f5 8.0-0
♗xc3 9.bxc3 0-0 leaves Black doing
well. ]
7...♗xc3
[ QUESTION: Isn't 7...♘xc3 8.bxc3 ♗xc3
possible too?
 Bronstein − Move
98 by Move
A N S W E R : C e r t a i n l y , b u t a f t e r 9.♖b1
Wh ite has always been consider ed to
have excellent compensation for the pawn.
The bishop on c3 will be forced to retreat
to a5, from where it will find it difficult
to return to active play, and White is ready
to advance in the centre with e4. The
black kingside is also distinctly devoid of
defensive pieces. With the best possible
defence, Black may be OK here, but his
game is not easy to play, and most top
players do not trust the position. ]
8.bxc3 0-0
[ QUESTION: Why not take the pawn with
8...♘xc3 ? This also allows Black to take
on f3 as well. Surely that is good for him?
ANSWER: Actually, no. His knight on c3
has trouble getting out. The line continues
9.♕c2 (the knight is trapped) ♗xf3 10.gxf3
♕g5+ 11.♔h1 ♕h5 and Black is already
playing only for a draw. However, instead
of acquiescing in the perpetual check,
White can play the more ambitious 12.♖g1!
♕xf3+ 13.♖g2 f5 (the only way to save
the knight, exploiting the fact that taking
on c3 leads to another perpetual) 14.♗a3
♘e4 15.♖f1 , threatening 16 Be2 and 17
Rxg7. White has a very strong initiative. A
famous game P.Keres-B.Spassky, 8th
matchgame, Riga 1965, saw the future
world champion crushed unmercifully from
this position: ♖g8 16.♗e2 ♕h3 17.f3 ♘f6
18.d5 ♔f7 19.e4 c5 20.♗b2 f4 21.e5
♘h5 22.♔g1 g6 23.♖g4 ♖d8 24.♗d3
♖g8 25.♖f2 and Black resigned, as his
queen is dropping off after 26 Bf1. Note
how Black's queen's rook and knight never
moved.
QUESTION: So does this mean Black's
entire set-up is just bad?
ANSWER: Actually, no, but he should
refrain from taking the pawn on c3.
Bronstein − Move by Move
Golombek does this with 8...0-0, but this
is also inaccurate. ]
[ M o d e r n t h e o r y r e c o m m e n d s 8...f5!
as best, when the main line is 9.d5! ♘c5!
10.♗c2 ♘ba6 and Black establishes his
blockade, albeit at the cost of his queen's
bishop and knight being at least
temporarily offside. Theory considers the
chances about equal. ]
9.♘e1! Once again exploiting the fact that
the c3-pawn is invulnerable, because the
knight cannot get out again. White frees the
path for his f-pawn, to expel the Black
knight.
QUESTION: But why is this so good? It
looks quite retrograde to put the knight back
on e1.
ANSWER: The point about this position is
the strategical battle for the square e4 and
the matter of activating the white bishop-pair.
As in many Nimzo lines, Black has
surrendered the two bishops and given White
a potentially powerful pawn centre, but has
doubled the c-pawns. If he can establish a
blockade and keep the position closed, the
bishops will not be able to become active,
and Black can take over the initiative. On the
other hand, if White can get his central
pawns moving and keep them mobile, and
can open the position for his bishops, then he
will stand better. Black would like to be able
to prevent the white pawn ever reaching e4,
in which case the white queen's bishop will
be stuck on a passive square, hemmed in by
the e3-pawn. If the bishop comes to a3,
Black wants to be able to play ...c5, again
shutting it out. As we will see, thanks to his
inaccurate 8th move, Black fails to
implement the blockade, and it is White's
strategy which triumphs.
f5 10.f3 ♘f6 11.a4! QUESTION: Why is
this good?
 Bronstein − Move
99 by Move
ANSWER: The point is to deny Black time to
play ...c5 and then ...Nc6.
[ Now 11.a4 would be met by c5 12.a5
with a nice queenside initiative. Black
therefore feels he has to stop the white a-
pawn advancing to a5, but in order to do so,
he must obstruct his own c-pawn. ]
11...♘c6 Now the c7-pawn is blocked, Black
will find it much harder to establish a central
blockade, and will also have always to
reckon with the thrust c4-c5, undoubling the
white c-pawns and trying to open the
position for the white bishops.
[ QUESTION: What if he stops the a-pawn
i n a d i f f e r e n t w a y , w i t h 11...a5
, keeping the chance to follow up with ...
c5 and ...Nc6?
ANSWER: But then a later ...c5 would
leave the b6-pawn very weak. Black also
wants to keep the a5-square free for his
knight, as he would like at some stage to
attack the c4-pawn with ...Na5 and ...Ba6. ]
12.e4 And this is the other downside of
putting the knight on c6. White has relieved
the blockade on e4 and got his central pawns
in motion.
QUESTION: It looks as though White is
getting what he wants?
ANSWER: Indeed. Black has been outplayed
over the past few moves and now has some
problems.
fxe4 13.fxe4 e5 Black must do something
to restrain the white central pawns, and this
is the logical way to do so. However, it does
weaken the square d5, which Bronstein
immediately exploits, but I still think 13...e5
is the only correct move.
[ Interestingly, Bronstein himself, in the
tournament book, comments that: "A less
brave and more cautious player would have
played 13...d6 here". There cannot have
been many occasions when Harry
Bronstein − Move by Move
Golombek was described as being brave or
lacking caution. Bronstein's point is
p r e s u m a b l y t h a t a f t e r 14.♘f3
( but I think instead, Bronstein would have
continued 14.♘c2 , keeping his options
open ) , B l a c k n o w p l a y s 14...e5
and the white knight is on the wrong circuit
to be able to exploit the d5-square. ]
14.♗g5 The bishop has found freedom and
now exerts unpleasant pressure on the black
position.
♕e7 Golombek's move is part of a subtle
idea, and reveals why he has avoided playing
...d6; he wants to keep the d6-square for his
queen.
[ The computer thinks Black has no
p r o b l e m s a f t e r 14...h6 , but I imagine
Golombek was afraid to weaken his
kingside in this way.
QUESTION: But where would the bishop
retreat to?
ANSWER: The mov e 15.♗h4 is quite
possible, since Black dare not try to grab a
pawn with g5 16.♗g3 exd4? in view of
17.e5 ♘h7 18.cxd4 , when the d4-pawn is
immune because of ♘xd4?? 19.♗xh7+ . ]
15.♘c2 ♕d6 QUESTION: What do you
think of this? It looks a bit strange.
ANSW ER: Yes, i t do es . Th e id ea i s t o
pressure the white centre, specifically the d4-
pawn, but tactically, the threats do not work,
and the queen is not great on d6.
16.♗h4! QUESTION: Why the exclamation
mark? Doesn't the move just lose the pawn
on d4?
[ ANSWER: No. The d4-pawn cannot be
taken, because after 16.♗h4 exd4 17.cxd4
♘xd4? there follows 18.♗g3 ♕c5 19.♗f2
and Black loses a piece. Meanwhile, 16
Bh4 sets up possibilities of Bg3, pinning
the black e-pawn, which could be very
awkward in some lines. ]
 Bronstein − 100
Move by Move
[ QUESTION: I notice White seems to be
avoiding the move 16.d5 . Why is this?
ANSWER: That would be just the
strategic concession Black is hoping for. As
we said earlier, with his two bishops,
White wishes to keep the position open and
his central pawns mobile. Playing d4-d5
would ossify the central pawn structure
and present Black with the chance to
implement the classic Nimzo-Indian
strategy against the doubled white c-pawns,
with ...Na5, ...Ba6, etc. Instead, White
keeps the three pawns abreast on c4, d4 and
e4, controlling a mass of central squares,
whilst he directs his knight towards d5. ]
16...♖ae8 Now there is a threat against the e4-
pawn after 17...exd4, so White defends by
pinning the e5-pawn.
17.♗g3 ♕e7 Unpinning e5 and so again
threatening to take on d4 and e4.
18.♘e3!
[ Once again dashing Black's hopes of
inducing the advance 18.d5? . Instead,
Bronstein directs his knight towards d5 and
f5. ]
18...d6 QUESTION: But why can't Black
take on d4?
[ ANSWER: After 18...exd4 19.♘f5 ♕d8
20.cxd4 ♘xe4 , there is the strong reply
21.♕g4! , threatening mate on g7 and
attacking the knight on e4. Black has no
choice but to defend with ♕g5 ( 21...♘g5?
22.h4 ♘e6 23.d5 wins for White)
, but then after 22.♕xg5 ♘xg5 23.♗xc7
White regains his pawn, and his two
bishops and active pieces give him a clear
advantage in the ending. 24 Nd6 is a threat,
among other things.
QUESTION: It is funny how what is
apparently a strategical battle, over
whether White can avoid closing the centre
w i t h d 5 , s e e ms t o d e p e n d o n t a c t i c a l
Bronstein − Move by Move
variations.
ANSWER: That is the whole essence of
chess. Strategy and tactics go hand in hand,
and good strategical play always depends
on accurate tactical control. Great
strategists, such as Capablanca, Petrosian,
Karpov, etc, were all superb tacticians,
able to calculate variations very quickly
and accurately, and this talent was the
reason they were able to execute their
strategic conceptions so effectively. This
is why the oft-quoted distinction between
'positional players' and 'tactical players' is
such a nonsense. Petrosian was every bit
as good a tactician as Tal, the only
difference was the precise use to which he
put his tactical prowess – he used it to
implement strategical manoeuvring plans,
whereas Tal used his to play mating
attacks. ]
19.♗h4! Bronstein's handling of the bishops
in this game is a model. By returning to the
pin on the black knight, he makes the threat
of Nd5 a highly unpleasant one for Black.
♘d8
[ B r o n s t e i n m e n t i o n s 19...♕d7
as the alternative, without indicating which
move he prefers (he comments only that
Black faces a "difficult choice"), but as
F r i t z s h o w s , a f t e r t h i s , s i m p l y 20.d5
(now this is effective, because it removes
the pressure against d4) ♘e7 21.♗xf6
♖xf6 22.♖xf6 gxf6 23.♕f3 ♖f8 24.♘g4
with 25 a5 to follow, gives White a large
advantage, with threats on both sides of
the board. ]
[ QUESTION: And what about taking on
d4? It does win a pawn, doesn't it?
ANSWER: Interestingly, Bronstein does
not mention this move, evidently assuming
that the threat of 20 Nd5 is too strong to
ignore. The computer confirms this to be
 Bronstein − 101
Move by Move
the case. After 19...exd4 20.♘d5 ♕d8
21.cxd4 ♘xd4 22.♘xf6+ gxf6 23.♕h5
the attack is too strong: for example, ♖e5
( 23...♕e7 24.♗xf6 ♖xf6 25.♖xf6 ♕xf6
26.♕xe8+ wins ) 24.♕h6 and f6 drops off
next move. The black king is too weak to
survive without ruinous material losses. ]
20.♘d5 ♗xd5 21.cxd5 QUESTION: So
what is the assessment?
ANSWER: Black has avoided an immediate
catastrophe, and his position is solid enough,
but strategically, he is in serious trouble.
His position is cramped, and all his pieces
passive, whilst White has tremendously
active pieces and the initiative in all areas
of the board. His bishops are all-powerful,
his pawn centre rock-solid, and he can play
on both the kingside and also the queenside,
after a4-a5. In fact, Black is just about
strategically lost.
c6 22.♕b3 ♔h8 23.♖ae1 h6 24.♕a3!!
QUESTION: What is the point of that? I
don't think it would have entered my head to
play this!
ANSWER: I sympathize with you. Bronstein
himself described it as "a difficult, but
useful move". In fact, I think it is the best
move of the game, by far. Black's 21st move
weakened his d6-pawn, and the white queen
now takes aim at that point. The true
effectiveness of the move becomes clear as
the game unfolds.
QUESTION: But doesn't this move lose the
d5-pawn?
ANSWER: Yes, but as we will see, White is
happy to trade that for the black centre pawn,
which will open the position completely for
the white bishops.
g5 25.♗g3 ♘d7 QUESTION: Why didn't
he take the pawn on d5? I thought that was
the point of playing 24...g5?
[ ANSWER: The trouble is, after 25...cxd5
Bronstein − Move by Move
26.exd5 ♘xd5 White just takes on e5:
27.dxe5 with a crushing position. White is
only too happy for the centre to open like
this, because in the resulting position, his
bishops will be devastating, especially
with the black king so exposed to the
elements. ]
[ QUESTION: OK, I see that. So maybe
Black should instead go after the bishop
with 25...♘h5 ?
ANSWER: A very sensible suggestion, but
Br o n s t e i n h a d f o r es e e n t h i s , a n d h a d
prepared a bone-crunching reply:
26.♗xe5+!! . Now we see another point of
his super-subtle 24th move. The following
variation is forced: dxe5 27.♕xe7 ♖xf1+
28.♖xf1 ♖xe7 29.♖f8+ ♔g7 30.♖xd8
and White has a winning endgame. ]
Golombek's move is more solid, just trying to
shore up his central structure, but the bishops
now embark on a de mo lition op eratio n
worthy of Fred Dibnah.
26.dxc6 ♘xc6 27.♗b5 As Bronstein points
out, this is what is known amongst chess
problemists as a 'half-pin': neither of the
black knights is actually pinned, but if either
one moves, the other will be.
♖xf1+
[ Thus, 27...♘db8? loses to 28.d5
when the knight on c6 is pinned. ]
[ A f t e r t h e i m m e d i a t e 27...♘cb8
, t h e r e p l y 28.♖f5 is deeply unpleasant,
with huge pressure against the black
centre. ]
28.♖xf1 ♘cb8 29.♗c4 Now 30 Rf7 is
threatened.
♖f8 30.♖xf8+ ♕xf8
[ 30...♘xf8? 31.dxe5 wins, once again
because of the pressure exerted along the
a3-f8 diagonal by the white queen. ]
31.dxe5 And here, too, the black centre
dissolves, thanks to the white queen's
 Bronstein − 102
Move by Move
pressure on the a3-f8 diagonal. The depth
and strength of Bronstein's 24th move is
very clear to see.
♘c5
[ 31...dxe5 32.♕xf8+ ♘xf8 33.♗xe5+
wins the knight at b8. ]
32.exd6 This position is worth comparing
with the one after White's 25th. The black
centre has done a real "here you see it, here
you don't" act. The devastating strength of
the white bishops is obvious.
♘xe4 33.d7! A nice combination to finish.
♘c5
[ The point of White's play is the variation
33...♘xd7 34.♕xf8+ ♘xf8 35.♗e5+!
♔h7 and now the sting in the tale: 36.♗d3
and the knight at e4 drops off. A
beautifully thematic variation, since the
whole game has been about the strength of
the white bishop-pair, and this position
illustrates it perfectly. ]
34.♗e5+ ♔h7 35.♗d3+!
[ 35.♗d3+ ♔g8 36.♕a2+ is curtains. An
absolute textbook on the exploitation of
the bishop-pair in the middlegame, and
replete with many of the delightful
combinational touches that characterize
Bronstein's best games. ]
1-0
35 E68
Zita,D
Bronstein,D
Moscow-Prague match, Moscow (Game 26)
[Steve Giddins]
1.c4 e5 The game starts out as an English
Opening, but soon transposes into the King's
Indian Defence.
2.♘c3 ♘f6 3.♘f3 d6 4.d4 ♘bd7 5.g3 g6
6.♗g2 ♗g7 7.0-0 0-0 This is a typical
Bronstein − Move by Move
King's Indian tabiya. Both sides have
developed their pieces within their own
camp.
QUESTION: The white bishop is on g2,
rather than e2, as is more often seen these
days. Is it worse there?
ANSWER: No, not at all. Although
variations such as Nf3 and Be2 and the
S ‫ن‬m i s c h w i t h f 2 - f 3 a r e p e r h a p s m o r e
common nowadays, the g3 system has always
been one of the most highly-respected set-
ups for White in this opening. It was for
many years a favourite with such KID-killers
as Korchnoi and Petrosian, and to this day,
many KID practitioners hate playing the
black side of this system.
QUESTION: What is the logic of putting the
bishop on g2?
A N S W E R : W e l l , t h e S ‫ن‬m i s c h b e i n g
something of an exception, White rarely gets
to attack the black king in this opening;
rather, Black is the one who usually becomes
active on the kingside, whilst White's play is
generally on the other flank. From that
viewpoint, it makes sense to have the bishop
pointing down the long diagonal, towards
the enemy queenside.
8.b3!?
[ Nowadays, the main line is 8.e4 c6 9.h3
. White's plan appears logical enough, but
as we shall see, the long diagonal is not
necessarily the best place for the white
queen's bishop. It is usually better placed
on e3, supporting the key d4-square. ]
8...♖e8 9.♗b2 c6 10.e4 exd4 11.♘xd4
QUESTION: I always have difficulty
understanding why Black makes the
exchange ...exd4 in such positions. Hasn't he
just surrendered his central outpost and left
himself a weak backward pawn on d6?
ANSWER: Well, you are not alone in finding
this difficult to understand. When Black first
 Bronstein − 103
Move by Move
started to play this way, most grandmasters
were extremely sceptical, for just the reasons
you give. It was games such as this and the
next, which opened people's eyes to the
viability of such an approach by Black.
QUESTION: So what is the point?
ANSWER: Well, firstly, the exchange opens
the long diagonal of Black's KID bishop, and
the half-open e-file for his rook on e8. So, he
has some pressure against e4 and d4. He can
augment the former with ...Nc5, which
exploits another advantage of the exchange
on d4, namely the freeing up of the c5-
square as an outpost for the black pieces, and
e5 is another possible outpost.
QUESTION: OK, but White also gets some
moves, too! What is to stop him just piling
up on the d-file and attacking the weakness
on d6?
ANSWER: Ah, well now you are getting to
the real secret of such positions. In his
famous book on the 1953 Zurich Candidates,
Bronstein dealt with this question head-on,
as it is crucial to the viability of Black's
play. On the face of it, it looks the simplest
thing in the world, just to bring a rook to the
d-file and then move the knight away from
d4, attacking d6. But, as Bronstein explains,
the d4-knight is actually a very important
piece and fulfils a number of key roles on
that square. It shields the long diagonal of
the enemy KID bishop on g7, it restrains the
possible enemy pawn breaks ...b7-b5 and ...
f7-f5, and it covers the e6-square, preventing
the enemy bishop from coming to that square
and attacking the c4-pawn. Consequently, it
is not so easy for the knight to move away,
and all the while it stays on d4, the d6-pawn
is not so easy to get at.
QUESTION: So the weakness of d6 is not
important?
ANSWER: It might be, if Black fails to make
Bronstein − Move by Move
use of his chances, but with accurate play by
Black, it is usually possible to generate
enough activity to prev ent White from
settling down to the exploitation of the
weakness. Black must remain active, and
make full use of all of his possible 'guerrilla
tactics', by harassing the e4 pawn,
threatening tactics along the a1-h8 and g1-a7
diagonals, using the half-open e-file, etc.
♕b6 12.♕d2?! QUESTION: What is wrong
with this? It looks very natural.
ANSWER: Maybe, but actually, the move
weakens the b3-pawn, of all things.
QUESTION: You are joking?
ANSWER: No, not at all. As we will see,
Black follows up with ...a5-a4 and takes on
b3, and then ...Nc5, when the b3-pawn will
be attacked.
[ K a s p a r o v p o i n t s o u t t h a t 12.♕c2
was better, defending both b3 and e4, with
the point that the knight on d4 is indirectly
defended, thanks to ♕xd4?? ( the other
tactic 12...♘xe4 also fails after 13.♘a4
followed by the capture on e4) 13.♘a4
trapping the queen. ]
12...♘c5 13.♖fe1 a5 This move is standard
in the King's Indian, and usually has the
primarily defensive function of preventing
the knight on c5 being driven away by b3-b4.
Here, however, that is not a threat anyway,
and ...a5 is instead a more aggressive move,
intending queenside play with ...a5-a4.
QUESTION: But White is still doing fine
here, isn't he? After all, he has more space
and the biggest weakness on the board is still
the d6-pawn.
ANSWER: Actually, I think White is
already slightly worse. As we have explained,
he cannot really generate any pressure on d6,
whilst Black's threats against e4, d4 and b3
give him the initiative. White is still holding
the position, but he needs to be careful.
 Bronstein − 104
Move by Move
14.♖ab1
[ K a s p a r o v s u g g e s t s 14.♘c2 a4
( but Fritz's 14...♘g4 looks more effective,
maintaining some initiative for Black)
15.b4 ♘e6 "with a complicated position". ]
14...a4 15.♗a1?! This looks solid enough,
but, as we will see, White is sleepwalking
into a brilliant combination.
[ The computer agrees with Kasparov, in
recommending 15.♘xa4 ♘xa4 16.bxa4
♕a5 , when Black is at least equal, but
probably no more than that. ]
15...axb3 16.axb3 ♘g4! This move is the
prelude to a brilliant combinative breakout.
The attentive reader will note that, as
indicated in the note to White's 8th move,
the normal set-up nowadays for White in this
line is to include the move h2-h3. It is
precisely to prevent moves such as Black's
last that h3 is usually played.
QUESTION: But what is going on? Why
can't the knight just be hit by 17 h3?
ANSWER: Ah well, read on!
17.h3? This natural response loses by force.
[ The only way to hold the white position
t o g e t h e r w a s 17.♖ed1 , reinforcing the
defence of the knight on d4. Then after
♘e5 18.h3 h5 (Kasparov), Black
maintains the initiative, but at least White
is still on the board. ]
It is worth looking at the position after 17 h3
mo r e c l o s el y an d s tu d y in g t h e v ar i o u s
geometric features, because they determine
the availability of Bronstein's combination.
One can detect various tactical motifs: the
bishop on g7 and queen on b6 both attack
(indirectly) the knight on d4; that knight
defends the b3-pawn; the rook on b1 also
defends the b3-pawn, but can be deflected by
the exchange sacrifice ...Rxa1; the knight on
g4 attacks f2 and a sacrifice on that square
would open the white king to a possible
Bronstein − Move by Move
check from the black queen; recapturing
with the king on f2 might allow a knight
check on d3, whilst Qxf2 might also allow ...
Nd3, hitting f2, e1 and d4. All in all, to the
experienced player, the position screams out
for a combination, but Black must find and
calculate the correct sequence.
17...♖xa1!
[ Bronstein himself points out that the
immediate 17...♘xf2? appears equally
good, until one sees the reply 18.♘a4!
. Then ♘xa4 19.bxa4 ♘xh3+ 20.♗xh3
♕c5 still gives Black enough for his piece,
but he is not actually winning, as he is
after the game continuation. ]
18.♖xa1 ♘xf2! QUESTION: Goodness!
Half the pieces on the board suddenly seem
to be hanging! What is going on?
ANSWER: Well, the tactical motifs outlined
in the note to Black's 17th now all come into
play.
19.♖e3
[ Firstly, with the white rook no longer on
the b-file, the move 19.♘a4? now makes
n o s e n s e , s i n c e a f t e r ♘xa4 20.bxa4
Black's queen is not attacked, and so he
c a n s i m p l y p l a y ♗xd4 with a winning
advantage. ]
[ Likewise, 19.♔xf2 ♘xb3 ]
[ and 19.♕xf2 ♘d3 are both disastrous for
White. ]
[ Moving the king off the g1-a7 diagonal,
and at the same time protecting the h3-
pawn, by 19.♔h2 , loses material after
♘cd3 , when e1 and d4 are both hanging. ]
Zita's move tries to block the g1-a7 diagonal
and protect the d3-square, but it does not
solve all the problems.
19...♘xh3+ 20.♔h2 ♘f2! The impudent
knight puts itself en prise again. Now it is
heading for g4, with devastating effect.
21.♖f3
 Bronstein − 105
Move by Move
[ 21.♗f3 is relatively best, but then simply
♘g4+ 22.♗xg4 ♗xg4 leaves ...Bh6
threatened, and 23.♖ee1 ♘xb3 24.♘xb3
♗xc3 25.♕xc3 ♕f2+ wins at once. ]
21...♘cxe4 22.♕f4
[ Or 22.♘xe4 ♘xe4 23.♕f4 f5
and the pin on the d4-knight costs yet
more material. ]
22...♘g4+ 23.♔h1 f5 Once again, the
deadly pin along the long black diagonal
wins further material.
QUESTION: The KID bishop certainly
proves its worth in this game!
ANSWER: Indeed. One can easily see why
this game was so influential. Who wouldn't
want to play the King's Indian after seeing
this game?
24.♘xe4 ♖xe4 25.♕xd6 ♖xd4 26.♕b8
♖d8 Black has two pieces and two extra
pawns for the rook, and his minor pieces all
exhibit deadly activity.
27.♖a8 ♗e5 28.♕a7 ♕b4 29.♕a2 ♕f8!
A nice geometrical touch. The queen
switches to h6, as White continues to be torn
apart on the dark squares.
30.♗h3 ♕h6
0-1
36 E68
Pachman,L
Bronstein,D
Moscow-Prague match, Moscow (Game 27)
[Steve Giddins]
1.d4 ♘f6 2.c4 d6 3.♘c3 e5 This Old
Indian move order was quite a popular way
to reach the King's Indian in the early days
of the system's popularity. The idea was to
avoid the S‫ن‬misch variation, which was
r e g ar d ed as v e r y d an g er o u s f o r B l ac k .
Subsequently, though, Black found various
Bronstein − Move by Move
improved methods against the S‫ن‬misch, and
the normal King's Indian move order became
more popular.
4.♘f3
[ QUESTION: Wouldn't White stand better
after 4.dxe5 dxe5 5.♕xd8+ ♔xd8
, with Black having lost castling rights?
ANSWER: Not really. Various games have
shown that Black can play ...c6, tuck his
king away on c7, and organize his pieces
quite well in this position. White does not
have any real advantage here. ]
4...♘bd7 5.g3 Now we head back to the
same line as in the previous game.
g6 6.♗g2 ♗g7 7.0-0 0-0 8.b3?!
Once again this mistaken development. As
we noted in the previous game, the bishop is
better placed on e3 in this line, keeping a
weather eye on the important d4- and f2-
squares.
♖e8 9.e4 exd4 10.♘xd4 ♘c5 11.♖e1 a5
12.♗b2 a4
[ 12...c6 was a perfectly good alternative,
not hurrying with the advance of the a-
pawn, but the text is also good. ]
13.♖c1?!
[ QUESTION: I am confused – I thought the
main point of playing ...a7-a5 in such
positions was to secure the position of the
knight on c5. Why doesn't White drive the
knight away with 13.b4 and at the same
time keep the a-file closed?
ANSWER: The problem is that Black has
the unpleasant reply a3! and after 14.♗c1
♘e6 , White's pieces are hanging rather
unpleasantly on the long diagonal. ]
[ For that reason, White hurries to move his
rook from the long diagonal, but Kasparov
points out that 13.♖b1! was a superior way
of doing this, keeping the bishop on b2
defended. The rook does not really do
anything much on c1, whilst the bishop on
 Bronstein − 106
Move by Move
b2 and the pawn on b3 may later suffer
from being undefended. ]
13...c6?! QUESTION: Why the punctuation?
Isn't this a typical idea in such positions?
[ ANSWER: Yes, but Kasparov points out
that it is not fully accurate at this point,
and that Black should first interpolate the
e x c h a n g e o n b 3 : 13...axb3 14.axb3
a n d o n l y t h e n c6 with a good game for
Black. ]
14.♗a1?! White misses his chance.
[ Now his rook has left a1, this was the
r i g h t t i m e f o r 14.b4 a3 15.♗a1
. The point – the bishop can now remain on
the long diagonal. After ♘e6 16.b5
White retains a small initiative. ]
14...axb3 15.axb3 ♕b6 Now play assumes
the characteristics of the previous game, and
Black has an ideal set-up. As before, his
pressure against b3, d4 and e4 prevents
White from turning his attention to the
exploitation of the potential weakness on d6.
16.h3 QUESTION: This is a standard
precaution in such lines, isn't it?
ANSWER: Yes, and we saw in the previous
game that White suffered through not having
played this move, as the black knight
suddenly jumped to g4 at a crucial moment.
QUESTION: So Pachman had learnt from
the previous game?
ANSWER: Well, no, actually that is the
funny thing – the present game was in fact
played four rounds earlier than the previous
one.
QUESTION: So Zita omitted h3 in an
attempt to improve on this game?
ANSWER: Probably, yes, but if so, he was
misguided. As we will see, in the present
game, the weakening of the white kingside is
exploited by ...h5-h4, and that may have
convinced Zita that White should keep his
pawn on h2, but even so, the importance of
Bronstein − Move by Move
guarding g4 is greater.
♘fd7 QUESTION: So what is the idea here?
This looks an odd retreat, shutting in his
queen's bishop.
ANSWER: Since the knight cannot come to
g4, Bronstein re-routes it via d7. This move
is a very standard one in this line – Black
unmasks his KID bishop, which is firing
down the long diagonal towards the rather
loose white knight on d4, whilst the black
knight prepares to swing round to e5, with
possible threats of jumping into d3.
QUESTION: You sound as though you
already prefer Black, yet White's position
still looks better – he has more space and his
development looks harmonious enough.
ANSWER: Appearances are deceptive. Black
is definitely somewhat better here, in my
opinion. White's position may look
superficially attractive, but, just as in the
last game, he has all manner of potential
tactical weaknesses: d4, b3, d3, e4, f2, etc.
Extra space is a good thing, if the opponent
has too little room in which to manoeuvre
his pieces comfortably, or if the side with the
space controls the initiative and can increase
the pressure gradually. But when the more
cramped side has adequate manoeuvring
room and the initiative, then the extra space
can easily become a liability.
It is a little like the military problems often
faced by large countries or far-flung empires
– it looks very nice to see so much of the
map bearing the colour of one's empire, but
all these far-flung places can be very
difficult to defend militarily, especially
against small bands of localized guerrillas.
This position is rather similar – the black
army resembles a guerrilla force, sniping
away at the various outlying portions of the
white position, whilst White's central
imperial army desperately tries to keep a lid
 Bronstein − 107
Move by Move
on everything.
17.♖b1 Admitting his error at move 13. The
threat of ...Ne5, with a possible fork on d3,
plus the weakness of the b3-pawn, convinces
White that his rook belongs on b1, after all.
♘f8?! QUESTION: I thought you said the
knight was coming to e5?
[ ANSWER: I did, and I think that is what
it should have done. Kasparov says
17...♘e5 is met by 18.♖e3 , but then the
c o m p u t e r ' s ♗h6 seems to keep up the
advantage.
QUESTION: But what about just 19.f4
then?
ANSWER: This loses at once ( instead,
W h i t e s h o u l d r e t r e a t w i t h 19.♖e2
, b u t a f t e r ♗d7 Black retains the
advantage ) , to the typical combination
19...♖xa1! 20.♖xa1 ♘xb3 . ]
18.♔h2?! QUESTION: I am surprised you
don't like this move. Surely it is 100%
natural, getting the king off the g1-a7
diagonal and preparing a later f2-f4
advance?
ANSWER: That is all true, but the move also
weakens f2 and encourages Black's next.
Kasparov gives a long note at this point,
explaining that part of White's difficulty here
is that almost any move he makes weakens
his position further.
QUESTION: A sort of 'zugzwang-lite'?
ANSWER: Yes. It is the great difficulty in
handling such positions as White. It requires
extraordinarily deep tactical penetration and
positional understanding, to be able to know
whether an apparently insignificant weakness
is going to prove serious.
[ The computer suggests the waiting move
18.♖e3 , which is probably best.
QUESTION: It looks a funny move!
ANSWER: Maybe, but it makes sense. The
g1-a7 diagonal is blocked, d3 is protected,
Bronstein − Move by Move
and an indirect extra defence of b3 is
added. ]
18...h5! An excellent move, further harassing
the white position. The advance of this pawn
to h4 will weaken the white kingside further,
wh ilst ...Nh7-g5 also becomes another
potential idea.
19.♖e2?! QUESTION: You don't seem to
like any of White's moves!
ANSWER: Well, I have great sympathy with
him. This is definitely a position where the
old Russian expression applies: "Byelim
dorog horoshiy soviet" – good advice is
expensive for White. It is really very
difficult to know what to recommend.
[ Instead, 19.h4?! radically stops the
advance of the h-pawn, but at the cost of
ceding g4, which Black can immediately
exploit with ♘h7 followed by ...Nf6-g4. ]
[ Q U E S T I O N : O K , h o w a b o u t 19.b4
, trying to simplify?
ANSWER: Unfortunately, that once again
walks into a combination: ♖xa1! 20.♖xa1
♕xb4 21.♖e3 ♕xc4 (Kasparov), and
Black has two pawns and a dominant dark-
squared bishop for the exchange – a clear
advantage. ]
[ QUESTION: So what should White do?
A N S W E R : K a s p a r o v g i v e s 19.♕d2
as relatively best, when he assesses h4
20.gxh4 ♘cd7 21.♖ed1 ♕c5 22.♔g1
♕h5 23.♕g5 ♕h8 as "unclear".
QUESTION: The black queen ends up on
h8?
ANSWER: Yes, Reti would have approved.
But the queen is very strong there,
supporting the KID bishop down the long
diagonal, as well as eyeing up the h4-pawn.
Pachman's move looks natural, defending
f2 and preparing to bring the rook to d2,
to shore up the knight on d4, but
unfortunately, it fails tactically. ]
 Bronstein − 108
Move by Move
19...h4 20.♖d2? Continuing the idea from
the previous move, but running into a
decisive tactical blow.
[ Kasparov gives 20.♖c2 as "a worse, but
still defensible position", ]
[ w h i l s t F r i t z s l i g h t l y p r e f e r s 20.b4
, although Black is still clearly better after
♘cd7 . ]
Now Bronstein strikes with a combination
extremely similar to that seen in the previous
game. The exchange sacrifice on a1 is again
the key, removing the bishop that is the
linchpin of the white position, whilst the
loose units on d4 and b3 are also a key
element in the combination.
20...♖xa1! 21.♖xa1 ♗xd4 22.♖xd4 ♘xb3
23.♖xd6
[ K a s p a r o v p o i n t s o u t t h a t 23.♖d2
is slightly more tenacious, preventing
Black's next, but even so, after simply
♘xa1 24.♕xa1 hxg3+ 25.fxg3 ♕b4
Black has an extra pawn and the white
pawn structure is wrecked. Black should
wrap this up without too much trouble. ]
23...♕xf2! QUESTION: Can't Black just take
on a1 first?
[ ANSWER: No! That would have fallen
into Pachman's trap. There would follow
23...♘xa1? 24.♘d5! turning the tables
completely. Now White is the one who
stands better after ♕xf2 25.♘f6+ ♔h8
26.♘xe8 hxg3+ 27.♔h1 ♘b3
( 27...♗xh3? 28.♕xa1+ is check ) 28.♖f6
, etc. ]
24.♖a2
[ Bronstein points out the attractive mate
after 24.♕xb3 hxg3+ ( the prosaic
24...♕xg3+ 25.♔h1 ♕xd6
is also winning) 25.♔h1 ♗xh3! 26.♖g1
♗xg2+ 27.♖xg2 ♕f1+ 28.♖g1 ♕h3# . ]
[ 24.♖d3 is possibly the best, but after
♘xa1 25.♕xa1 ♘d7 (Bronstein), there is
Bronstein − Move by Move
no hope for White, as the computer
confirms. ]
24...♕xg3+ 25.♔h1 ♕xc3 26.♖a3 ♗xh3!
Winning nicely. Kasparov gleefully points
out that, just as 20...Rxa1! was the first move
by the black queen's rook, so this decisive
blow is the first move of his queen's bishop.
[ The computer (and Kasparov) adds that
26...♖xe4! a l s o w i n s : 27.♗xe4 ♕xh3+
28.♔g1 ♕e3+ 29.♔h2 ♕f4+ 30.♔g1
♗g4 31.♕xb3 ♕xd6 , etc. ]
QUESTION: It is amazing how it all fits
together tactically.
ANSWER: Well, that is the greatness of a
player such as Bronstein. But the whole
thing is of course based on his superb
positional build-up, and the looseness of the
various white units.
27.♖xb3 ♗xg2+ 28.♔xg2 ♕xc4 29.♖d4
♕e6 30.♖xb7 ♖a8 31.♕e2 h3+
[ There is no defence after 31...h3+ 32.♔g1
♕e5 followed by ...Ne6. It is safe to say
that these last two games did more than
any others to popularize the King's Indian
Defence, which went on to become just
about the single most popular defence to 1
d4, over the next few decades. ]
0-1
37 E69
Reshevsky,S
Bronstein,D
Zurich Candidates (Game 28) 1953
[Steve Giddins]
1.d4 ♘f6 2.c4 g6 3.g3 ♗g7 4.♗g2 0-0
5.♘c3 d6 6.♘f3 ♘bd7 7.0-0 e5 8.e4 ♖e8
9.h3 QUESTION: The same variation as in
the previous two games?
ANSWER: Yes, but here we see how theory
had advanced in the intervening seven years.
 Bronstein − 109
Move by Move
Rath er than b3 and Bb2, which proved
unsuccessful in the earlier games, Reshevsky
adopts a plan, which is still the main line to
this day. The early h3 covers the vital g4-
square, preventing the black knight
manoeuvre ...Ng4-e5, and also preparing to
place his queen's bishop on e3, from where it
defends the knight on d4 and discourages the
black queen from occupying b6, as it did in
the two previous games.
exd4 10.♘xd4 ♘c5 11.♖e1 a5 12.♕c2
[ The main line of development, although
K a s p a r o v s u g g e s t s t h a t 12.♖b1
is more useful, intending to meet a4 with
13.b4 axb3 14.axb3 c6 15.b4 ♘e6
16.♘de2 when White retains some
advantage.
QUESTION: Why is he better here, when
he was not in the two previous games?
ANSWER: Because he is better organized
here. The black knight has been driven to
a less-than-ideal post on e6, where it
blocks the bishop on c8 and rook on e8.
White is now quite well-placed to follow
up with Be3, Qd2 and Rbd1, when he really
will have some pressure against the
backward d6-pawn.
QUESTION: The differences between this
and the previous two games don't seem
that great, yet the assessment is reversed!
ANSWER: Well, chess is a subtle game.
In general, these positions with ...exd4
ar e v er y d o u b l e- e d g ed . T h e f ir s t tw o
games showed an ideal scenario for Black,
where White was never able to generate
any pressure against d6, whilst White's
own weaknesses were under constant
pressure. But it only takes a few subtle
changes to the position for the roles to be
reversed – if Black's 'guerrilla tactics' are
n e u t r al i z e d , t h e n t h e w ea k n e s s o f d 6
really will become significant. This is
Bronstein − Move by Move
why such positions require extremely
precise handling by Black. It is not a set-
up for those who just like to be able to
play natural, common-sense moves – Black
needs to show really deep positional and
tactical insight. ]
12...c6 13.♗e3 ♘fd7 14.♖ad1 a4
QUESTION: So what is happening now?
White is fully mobilized, whereas most of
Black's pieces are on the back rank. Surely
White is better?
ANSWER: Possibly slightly, but not much.
Chances are fairly balanced. It is true that
Black has much less direct pressure than in
the preceding examples, but he still has some
ideas. He can bring his queen to a5 and his
knight to e5, hoping to pressure the c4-pawn.
If White defends it with b2-b3, then an
exchange of pawns on b3 will open the a-file
for Black, and a later ...Qb4 may exert some
pressure on the pawn. By these means, Black
hopes to prevent White settling down to the
exploitation of d6.
15.♘de2?! QUESTION: Why don't you like
this move? Isn't it logical to unmask the
attack on d6?
ANSWER: The idea is tempting in itself, but,
as we will see, it does not prove very
effective.
QUES TI ON: S o wh at sh ou ld Wh it e d o
instead?
ANSWER: A good question!
[ This was something of a 'hot' position in
the Zurich tournament, and was reached
several times. In Stahlberg-Boleslavsky,
W h i t e t r i e d 15.f4 , stopping the black
knight coming to e5.
QUESTION: That seems very logical, in
view of what you said about Black's plan
to attack c4.
ANSWER: Yes, but the move also weakens
the white e4-pawn. Play continued ♕a5
 Bronstein − 110
Move by Move
16.♗f2 , and now ♘b6 , hassling the c4-
pawn from another direction. After 17.♗f1
♗d7 Black followed up with ...Rad8 and
stood fine. ]
[ 15.♖e2 was tried by Averbakh, aiming to
double on the d-file, but after ♕a5 16.♖ed2
♘e5 17.b3 axb3 18.axb3
, Black proved to have adequate play.
This position is a good example of what
we discussed in Game 26, regarding the
position of the white knight on d4. Here we
see that White is all ready just to move
that knight away, uncovering a double
rook attack on d6, but the pressure against
b3 prevents the knight from leaving its
post. ]
[ QUESTION: So, what should White play
then?
ANSWER: Kasparov suggests the
interesting 15.b4!? , challenging the black
k n i g h t . A f t e r axb3 16.axb3
White continues to try to force the advance
of his b-pawn: 17.♖b1 . He claims some
i n i t i a t i v e f o r W h i t e a f t e r ♘a6?!
( but the computer does not like this at all
f o r B l a c k , a n d p r e f e r s 17...♕b4
, radically preventing the advance of the b-
pawn.
QUESTION: That looks a bit dangerous
after 18.♗d2 doesn't it? 19 Nd5 is a threat.
ANSWER: Indeed, but the silicon monster
assures us there is nothing to worry about
after ♕a3 when White has nothing better
than a repetition with 19.♗e3 ♕b4 , etc.
QUESTION: So, the overall verdict is one
of equality, is it?
ANSWER: Yes, I think that is fair. But
the variations show how precise the two
players must be, to maintain that balance
– even a small inaccuracy can be enough
to leave a player suffering, especially
Black. )]
Bronstein − Move by Move
15...♕a5! 16.♗f1 This looks rather
retrograde, and seriously weakens f3, but
Black was threatening ...Ne5 and/or ...Qb4,
attacking the pawn on c4, which cannot now
be defended by b2-b3. Note how White is
already missing his knight on d4, which
♕b4 would have been defending b3.
QUESTION: Huh? Isn't d6 hanging?
[ ANSWER: It turns out that after
16.♖xd6?! ♘e5 Black is threatening ...
Nxc4.
QUESTION: OK, so we defend it with
17.b3 .
ANSWER: Yes, and now Black reveals his
other threat: axb3 18.axb3 ♗xh3! .
QUESTION: But...
ANSWER: After 19.♗xh3 ♘f3+ 20.♔f1
♘xe1 21.♔xe1 White has two pieces for
the rook, but now follows ♘xe4 22.♖d3
♕a1+ and Black is winning: for example,
♕a5 23.♘b1 ♖a2 24.♕d1 ♗f6 (covering d8)
and Black's initiative is too strong.
QUESTION: Lightning from a clear sky!
, ANSWER: Yes, and absolutely typical of
what happens in these structures, when they
are handled well by such a great player as
Bronstein. There are all manner of hidden
tactical chances and Black continually
manages to keep his d6-pawn alive by
exploiting them. Matthew Sadler once used
the analogy that playing White in such
positions is a bit like trying to squash a
jelly – you press down on one bit and it
pops up elsewhere. ]
16...♘e5 Black keeps playing dynamically,
pressuring c4 and threatening ...Nf3+.
[ By contrast, a passive defence of d6 with
16...♗f8? would be a clear mistake, and is
almost always a bad sign in such
positions. ]
17.♘d4 The knight returns to d4, just two
moves after leaving it – another confirmation
 Bronstein − 111
Move by Move
of what we have said earlier, that it is quite
hard for White to do without the knight on
this square. Now d6 is no longer attacked, but,
as Kasparov points out, Reshevsky was
probably hoping this would only be
temporary, and that he would be able to
drive away the black knight with f4 and then
return to the task of harassing d6. However,
Black continues his guerrilla tactics.
a3! QUESTION: That is a bit of a surprise.
Doesn't this allow White to defend the c4-
pawn securely with b3, without the a-file
being opened or b3 itself becoming an object
of attack?
ANSWER: True, and it is noteworthy that
the computer does not particularly like this
move. But Kasparov awards it an
exclamation mark, and I agree. Effectively,
Black is trading positional advantages – yes,
he gives up his pressure against c4 and b3,
but in return, he establishes a blockade on
the dark squares c5 and b4, which will both
be occupied by the black knights. In addition,
the a3-pawn will be a trump card for Black
in any endgame.
18.f4 ♘ed7 19.b3 ♘a6 QUESTION:
Changing of the guard?
ANSWER: Indeed! This knight is heading to
b4, and its colleague will take over the c5-
square. Note too, how the moves ...a3 and b3
have weakened the long diagonal. Now the
white knight on d4 has to worry about its
colleague on c3, which may be attacked if it
moves away from d4.
QUESTION: So White still can't move his
knight from d4, to attack d6?
ANSWER: No, or at least, it is not entirely
simple.
20.♗f2 ♘dc5 21.♖e3 QUESTION: What is
this for? It looks a bit clumsy.
ANS WER: I t d o es , but Wh ite needs t o
defend the c3-knight (see the previous note),
Bronstein − Move by Move
so his queen is free to move away after ...
Nb4.
♘b4 22.♕e2 ♗d7 QUESTION: How do
you assess the position? White seems to
have been hassled a lot over the last few
moves, but now things are stable, aren't they?
ANSWER: Yes, that is true, but White still
has to find something to do. Black intends ...
Re7 and ...Rae8. What is White to do?
QUESTION: Well, I would like to attack the
d6-pawn!
[ ANSWER: Naturally, but you will note
that, yet again, the knight on d4 cannot
m o v e a w a y : 22...♗d7 23.♘f3?? ( or
23.♘c2?? ♗xc3 24.♖xc3 ♘xe4 )
23...♗xc3 24.♖xc3 ♘xe4 wins. ]
23.e5?!
[ White does not have anything particularly
c o n s t r u c t i v e t o d o , b u t t h e c o mp u t e r
suggests he should just wait, with a move
such as 23.♗g2 ]
[ or 23.♕f3 , when chances would be equal.
Instead, Reshevsky decides to get rid of his
e4 weakness, by trading it for Black's
weakness on d6. However, this backfires. ]
23...dxe5 24.fxe5 ♖ad8
[ T h e c o m p u t e r t h i n k s 24...♕c7
is even more accurate, but even after the
text, the assessment of the position has
changed.
QUESTION: Why?
ANSWER: Because Black now simply
stands better, as the pawn on e5 is weak,
and White has no compensation for this. ]
25.g4 QUESTION: That looks rather
weakening!
ANSWER: It is, but White needs to get his
bishop to h4 or g3.
♘e6 26.♗h4 ♘xd4 27.♖xd4 ♕c5!
As Bronstein notes, this is a key move.
Black must prevent the enemy knight
reaching f6.
 Bronstein − 112
Move by Move
[ And, of course, not 27...♗xe5?? 28.♖xe5
♖xe5 29.♗xd8 , winning for White. ]
28.♖de4 ♗h6! 29.♔h1
[ 29.♗xd8 ♖xd8 30.♕f2 ♗xe3 31.♖xe3
♕d4 32.♘e4 ♘xa2 is no better, ]
[ whilst 29.g5 ♗g7 is even worse, as the
black queen's bishop comes to f5 next
move. ]
29...♗e6 30.g5
[ Kasparov describes this as a desperate
attempt to complicate, since the variation
30.♗xd8 ♖xd8 31.♖g3 (the point of
White's 29th move) ♖d2 32.♕e1 ♘c2
33.♕b1 ♕f2 is hopeless. ]
30...♗g7 31.♖f4 ♗f5 32.♘e4 Setting a very
nasty trap.
♗xe4+!
[ 32...♕xe5?? loses to the vicious trick
33.♖xf5! gxf5 ( or 33...♕xf5 34.♘f6+
♗xf6 35.♖xe8+ ♖xe8 36.♕xe8+ ♔g7
37.gxf6+ ) 34.♘f6+ ♗xf6 35.gxf6!!
. The whole point. There is no defence
against Qg2+. ]
33.♖fxe4 ♘a6 QUESTION: Why is this
strong?
ANSWER: The knight heads to the
blockading square e6. Knights are well-
known for being the ideal blockader.
QUESTION: OK, but doesn't this allow
White to offload his weak pawn?
ANSWER: Yes, but, as will become apparent,
this does not solve his problems.
34.e6 fxe6 35.♖xe6 QUESTION: So how
do you assess this position?
ANSWER: Black has a clear advantage.
Bronstein himself (who turned down a draw
offer in this position) identifies three
factors: the exposed white king, the
superiority of the knight over the white
bishops, and the strength of the pawn on a3.
The last factor was something we mentioned
in the note to Black's 17th move.
Bronstein − Move by Move
♖f8?! QUESTION: Why the punctuation?
ANSWER: This is an interesting moment.
Both Bronstein and Kasparov award Black's
last move an exclamation mark, but Fritz
claims it is a mistake, after which the
chances are equal.
[ The computer prefers 35...♖xe6 36.♖xe6
♘b4 with an advantage. Note that
37.♖e8+? is then a mistake, since after
♖xe8 38.♕xe8+ ♕f8 White cannot defend
both f1 and a2. ]
36.♖e7?
[ Fritz claims equal chances after 36.♗g2
which Kasparov and Bronstein both ignore.
For example, ♘b4 37.♖e7 ♕d4 ( or
37...♗d4? 38.♖f3 ♖xf3 39.♗xf3
and the th reat o f Bg4-e6 is better for
White ) 38.♖xb7 ♕xh4 39.♖xg7+ ♔xg7
40.♖e7+ ♖f7 41.♕e5+ ♔g8 42.♖xf7
♔xf7 43.♕f6+ ♔e8 44.♕e6+
, forcing perpetual check. ]
36...♗d4 37.♖3e6 Now we see the difference
with the last note: White's pieces are hanging.
[ 37.♖f3?? is impossible because of ♖xf3
38.♕xf3 ♕xe7 , whilst the text leads to
problems down the f-file. ]
[ 37.♖3e4 is similar to the game: ♕f5
38.♗g2 ♘c5 , and now 39.♖g4 ♕b1+
40.♔h2 ♕g1+ 41.♔g3 ♗f2+
forces mate. ]
37...♕f5! 38.♖e8
[ 38.♗g2 ♘c5 traps the rook on e6 in broad
daylight. ]
38...♘c5 39.♖xd8
[ 39.♖6e7 is elegantly refuted by ♖dxe8
40.♖xe8 ♘xb3! – another illustration of
the strength of the a3-pawn. ]
39...♘xe6 40.♖xf8+ ♔xf8 QUESTION:
Wouldn't the game have been adjourned
around here?
ANSWER: Yes, it was, with Reshevsky
sealing his next move. His position is almost
 Bronstein − 113
Move by Move
certainly lost, as the g5-pawn is weak, the
white king exposed, and the a3-pawn still a
permanent threat. However, Black's king is
also exposed, and great care is required in
converting the advantage.
41.♗g3 QUESTION: Giving up the g5-pawn
at once?
ANSWER: Yes, Reshevsky decides to
jettison the pawn immediately, in search of
counterplay. There is already a trap.
♕xg5!
[ 41...♘xg5? 42.♗d6+ ♔g7 43.♕e7+
♔h6 44.♕f8+ ♕xf8 45.♗xf8+ ♔h5
46.♗xa3 is an immediate draw. ]
42.♕xe6 ♕xg3 43.♕c8+ ♔e7 44.♕g4
QUESTION: Why not take b7 with check?
[ ANSWER: Unfortunately, this is
i m p o s s i b l e , a s a f t e r 44.♕xb7+ ♔d8
White runs out of checks within a few
moves, and meanwhile, cannot stop the
mating threat: 45.♕a8+ ( or 45.♗g2 ♕e1+
46.♔h2 ♗e5# ) 45...♔c7! 46.♕a5+ ♗b6
.]
44...♕c3 QUESTION: So Black is a pawn up,
but we have opposite-coloured bishops. Is
that not a big drawing factor?
ANSWER: It is a drawing factor when there
are no other pieces on the board, but with
queens on, as here, the tendency is if
anything, the opposite – the possibility of a
attack, concentrated on the same colour
squares as one's own bishop, is extremely
dangerous, since the opponent effectively
has one fewer piece with which to defend
those squares. White's only drawing chances
are either perpetual check, or else to
exchange queens at a favourable moment,
when the pure bishop ending will be drawn.
However, the latter is not easy, as the
vulnerability of the a2-pawn to the black
king means that some of the bishop endings
will be lost for White.
Bronstein − Move by Move
45.♔g2 ♕b2+ 46.♕e2+ ♔d6
[ The position after 46...♕xe2+? 47.♗xe2
would be drawn, but instead, Black exploits
the fact that White cannot himself exchange
on b2, as the black king is too close. ]
47.♔f3
[ Kasparov quotes an old analysis by Euwe:
47.♕xb2? axb2 48.♗d3 ♔e5 49.♔f3 ♗c5
50.♗b1 ♔f6 51.♔e4 h5 52.♗c2 g5
53.♗b1 ♗d6 54.♗c2 g4 55.hxg4 hxg4
56.♗b1 g3 57.♔f3 ♗f4 and, with the
white king tied down to covering the g-
pawn, the black monarch just marches in on
the queenside. ]
47...♗c5 48.♔e4 ♕d4+ 49.♔f3 ♕f6+
50.♔g2 ♔c7 51.♕f3 ♕b2+ 52.♕e2 ♕d4
53.♔f3 h5 QUESTION: It looks as though
Black has been groping for the right plan?
A N S W E R : W i t h a d j o u r n me n t a n a l y s i s
behind him, I rather doubt that. It was
always a feature of Soviet technique not to
rush such positions and to take one's time to
strengthen the position to the utmost, all the
while the opponent has no counterplay. Now,
at any rate, Black commences the decisive
advance of the kingside pawns. This will
lead either to a mating attack or, after a
queen exchange, to a winning opposite-
bishop ending, similar to that analysed in the
note to White's 47th.
54.♔g2 g5 55.♔g3 ♕f4+ 56.♔g2 g4
57.hxg4 hxg4 58.♔h1 ♔b6 59.♔g2 ♔c7
60.♔h1 ♗d6
[ In his analysis, Bronstein had found a neat
zugzwang finish, but the computer points
out the much more prosaic and immediate
60...♕d4 61.♕g2 ♗d6 , and there is no
defence to mate after ...Qh8. ]
61.♔g1 ♔b6 62.♕g2 ♗c5+ 63.♔h1 ♕h6+
64.♕h2 ♕e3 65.b4 ♗d4
[ White is in zugzwang and lost on time. If
65...♗d4 66.♕g2 g3 67.c5+ ♔a7
 Bronstein − 114
Move by Move
. As well as the mere fact that Bronstein
o u t p l a y e d s u c h a n i mm e n s e l y s t r o n g
opponent as Black, this game is a nice
counterpart to the previous two. It shows
that, even if White plays much better and
av o i d s t h e s o r t o f t a c t i ca l s t o r m t h a t
engulfed him in Games 26 and 27, the
black position is fully viable, if handled
with sufficient expertise. ]
0-1
38 E81
Pein,M
Bronstein,D
London League (Game 29) 1997
[Steve Giddins]
This game was another from Bronstein's
appearances for Charlton Chess Club in
evening league chess.
1.d4 ♘f6 2.c4 g6 3.♘c3 ♗g7 4.e4 d6
5.f3 The S‫ن‬misch Variation.
QUESTION: So what is the idea?
ANSWER: White strengthens his centre and
prepares to place his bishop on e3, with the
likely follow up Qd2 and 0-0-0, followed by
attacking on the kingside with Bh6, g4 and
h4. This is one of the few lines of the KID
where White can hope to develop a direct
kingside attack.
QUESTION: Sounds good!
ANSWER: Well, it has always been one of
White's most popular weapons against the
KID, but the move does have its downsides,
as well. The main one is that White
experiences some problems with the
development of his kingside. The pawn on f3
deprives the g1-knight of its natural square,
and putting the knight on e2 is liable to
block in the light-squared bishop.
QUESTION: But why not just play Bd3 and
Bronstein − Move by Move
then Nge2?
ANSWER: The trouble with that is that the
square d4 can become a little tender. The
other main objection to the S‫ن‬misch is that
the move f2-f3 weakens the central dark
squares, especially e3, and the g1-a7 diagonal.
Black's counterplay is often directed against
these targets.
0-0 6.♘ge2
[ QUESTION: That looks a rather clumsy
move. Why not just 6.♗e3 , which you said
was the main idea?
ANSWER: For many years, the latter was
indeed almost the only move played here,
but in recent years, this has fallen out of
favour somewhat, because of a radical
method of counterplay: c5! .
QUESTION: But surely that just loses a
pawn?
ANSWER: Yes, it sacrifices one, and
allows the queens off to boot. But various
games played from the late 1980s onwards,
by players such as Gelfand and others,
showed that, surprising as it may seem at
first sight, Black has decent counterplay
after 7.dxc5 dxc5 8.♕xd8 ♖xd8 9.♗xc5
♘c6 .
QUESTION: It does not look convincing to
me!
ANSWER: No, I am not surprised. I felt the
same way for quite a time, and most other
players also have trouble convincing
themselves that it is really good. But the
more one studies the position, the more
one realizes that th e weakness of the
central dark squares d4, e3, etc, combined
with the log-jam of white pieces on the
kingside, gives Black good counterplay.
This is not the place to go into the
theoretical details, but if you consult any
good recent work on the King's Indian, you
will begin to see what I mean. ]
 Bronstein − 115
Move by Move
[ QUESTION: OK, accepting what you say
at face value for the moment, how does
6.♘ge2 help White? Now he is not even
winning a pawn after c5 .
ANSWER: No, indeed, but he intends to
meet this with 7.d5 , when he argues that
he has a superior Benoni than would be the
case with the knight on g1 and the bishop
on e3. Once again, it is all fairly subtle
modern theory, and is not really relevant
to our present game, to which we now
return. ]
6...e5 Bronstein was a much more classical
KID player than many modern exponents of
the opening, and he generally preferred ...e5
against the S‫ن‬misch, in most variations.
7.♗g5 This is another point of White's
choice at move 6. By delaying the
development of the bishop until Black has
advanced his e-pawn, White is now able to
bring the bishop to this more active square
than e3. The pin on the knight creates a few
problems for Black.
c6
[ QUESTION: Can't Black just kick the
bishop with 7...h6 ?
ANSWER: He could, and this is probably
the typical club player's reaction, but it
would be playing into White's hands by
weakening the black kingside. One must
remember that White is still in a position
to castle long and play for the attack on the
kingside, so such a weakening move as ...
h6 could really come back to haunt Black.
After 8.♗e3 followed by Qd2 (with tempo)
and then 0-0-0 and h4 plus g4, White
would have a very dangerous attack. ]
8.♕d2 ♘bd7 9.0-0-0 White continues
consistently, tucking his king away on the
queenside and preparing for a full-blooded
struggle. One must admire Malcolm Pein for
his approach here – many players, facing the
Bronstein − Move by Move
great David Bronstein, might have been
tempted to wimp out and play for a draw as
White, but Malcolm has never been that type
of player. He seizes the chance to take on his
great opponent in a sharp and double-edged
fight. And quite right, too, regardless of the
eventual result – after all, how often does
one get the chance to play against a
challenger for the world championship?
♕a5 10.♗h6?! Certainly consistent, but
actually a novelty, according to the database.
[ In almost all the practical tests of this
position (and there are not many), White
has preferred the prophylactic 10.♔b1
here, securing his a2-pawn and potentially
setting up Nd5 tactics against the
unguarded black queen.
QUESTION: But can Black continue as in
t h e g a m e w i t h b5 and be a tempo up? I
thought in such opposite-castling positions,
speed was of the essence?
ANSWER: It is, but the rule of 'more haste,
less speed' also sometimes applies. The
problem with 10...b5 in this position is that
W h i t e h a s t h e s t r o n g r e p l y 11.♘d5!
, when after the almost forced line ♕xd2
12.♘e7+ ♔h8 13.♖xd2 ♗b7 14.cxb5
cxb5 15.♘c3 he would have a nice
advantage. The b5-pawn is exposed, d5 is
weak and White is well placed to continue
with Be2 and Rhd1 (or Rc1), with the
advantage. ]
10...b5?!
[ It looks even stronger to interpolate
10...♗xh6! 11.♕xh6 , dragging the white
queen away, and only then to strike out
with b5! . Now White has some problems,
because 12...b4 is a threat, winning the a-
pawn, and this is not easy to meet:
12.♔b1?? b4 wins the knight, a line which
shows one of the problems of having the
king's knight on e2. ]
 Bronstein − 116
Move by Move
11.♗xg7 ♔xg7 12.♘g3
[ 12.♔b1?? is still impossible because of b4
, so White cannot protect his a2-pawn. ]
[ QUESTION: I suppose 12.a3 would be
worse still?
ANSWER: Well, that is actually not so
c l e a r . I t i s v e r y u n a t t r a c t i v e a f t e r b4
13.♘a2 c5 , but may not be as bad as
things should have been in the game. ]
12...b4 13.♘b1 exd4 QUESTION: Why
didn't he take on a2? I thought that was the
whole point?
[ ANSWER: 13...♕xa2 14.♕xb4
is not entirely clear, as d6 is hanging, and
♖b8 15.♕a3 forces off the queens. Black
prefers to keep the ladies on the board
and play for the counterattack. ]
14.♕xd4 ♕g5+?! QUESTION: Why didn't
he just take on a2?
[ ANSWER: Objectively, that is somewhat
better for Black, indeed. According to
Malcolm Pein, annotating the game in his
'Daily Telegraph' column at the time,
B r o n s t e i n d i s m i s s e d 14...♕xa2
as "skittles" (!), but the text is a step in
the wrong direction. ]
[ Fritz offers as best the move 14...♖d8
, which is not easy to find.
QUESTION: It just steps into a pin, doesn't
it?
ANSWER: Indeed, which is why it is
c o u n t e r i n t u i t i v e . B u t a f t e r 15.♘d2
( 15.♕xd6 ♗a6! threatens 16...Ne5, which
is very nasty for White) 15...♘c5 16.♔b1
♘e6 17.♕f2 ♕e5 Black has a nice plus,
with his grip on the central dark squares.
This is one of the positional issues with
the plan of exchanging dark-squared
bishops – as we have already pointed out,
the characteristic S‫ن‬misch move 5 f3
rather weakens the central dark squares in
White's camp, and so, from a positional
Bronstein − Move by Move
viewpoint, exchanging dark-squared
bishops is a risky business.
QUESTION: But the KID bishop is a key
defender of the black king, so it has to go,
doesn't it?
ANSWER: It does if White is hoping to
attack the kingside, yes. But the downside
is that, if the attack does not get going
rapidly, the central dark-square weaknesses
can come back to haunt White. ]
15.♘d2 ♘e5 16.♘e2?! White is tempted by
the idea of threatening f3-f4, but this knight
move is rather clumsy, as the knight returns
to the square in front of the still-undeveloped
bishop on f1.
[ Objectively, White should probably grab
t h e o f f e r e d p a w n w i t h 16.♕xd6
, when it is not clear Black really has
enough, although the position is still
quite complicated.
QUESTION: There seem to be quite a
few inaccuracies on both sides, around
here!
ANS WER: That is true, but on e mus t
remember that the game was played in a
L o n d o n L e a g u e e v e n i n g ma t c h . S u c h
games, played after a day at work, and at
a fast time-limit (I think this would
probably have been played at 35 moves in
75 minutes, or something similar) tend to
be dogfights, rather than models of
precision. What redeems the game, as far
as inclusion in this book is concerned, is
the final combinative finish. ]
16...c5 17.h4? Pein continues to seek ways
to sharpen the battle, and initiates a tactical
sequence, but objectively, it is unfavourable
for White.
[ He should probably settle for 17.♕xd6
♘xc4 18.♕f4 , when after the further
moves ♕xf4 19.♘xf4 ♘e3 20.♖e1 ♘xf1
21.♖hxf1 , chances are about equal in the
 Bronstein − 117
Move by Move

ending. ] 22.♗g2? Walking into the main line.

17...♕xg2! QUESTION: A desperado! [ Objectively, White should try 22.♔c2
ANSWER: Yes, and there are more to come... , when he threatens 23 Bg2 Qh2 24 Nf1.
18.♕xe5 Anything you can do... B l a c k w o u l d t h e r e f o r e c o n t i n u e ♗xe2
♕xh1 I'll have the exchange, thank you! 23.♗xe2 ♕h3 , b u t t h e n 24.♖g1
19.♕g3 The point. threatens 25 Bg4, as well as 25 Bxh5, so
QUESTION: Hmm. What on earth is going Black would have nothing better than ♘xf4
on here? 25.♕xf4 ( not 25.♗g4?? ♕d3+ ) 25...♕e6
ANSWER: White's 17th was effectively an with a position where he has rook and two
exchange sacrifice, in return for which he pawns for two minor pieces. Chances
has trapped the black queen on h1. The would be about equal. ]
threat is 20 Bg2, and it is not immediately 22...♘xf4! The brilliant point. The threat of a
obvious how Black can deal with this. But fork on d3 regains the queen.
unfortunately for Pein, there is an old saying: 23.♕xf4
"If you sup with the Devil, you need a long [ If 23.♖xh1 ♘d3+ 24.♔c2 ♘xf2
spoon!". That was certainly the case when it and Black is winning comfortably. Pein
came to trading tactics with David Ionovich, tries something else, but the black queen
and it turns out that the position favours is not getting trapped. ]
Black after all. 23...♕xg2 Another piece hangs on e2. It is all
♘h5 20.♕f2 over.
[ 20.♕xd6? ♕xh4 allows the black queen 24.♖g1 ♕xe2 25.♖xg4 ♖ae8 26.h5 ♖e5
to escape with booty intact, and so makes 0-1
no sense. ]
20...♔g8 21.f4 QUESTION: Can't he trap
the queen at once? 39 A17
[ ANSWER: The immediate 21.♗g2 ♕h2 Langeweg,K
(threatening 22...Qe5) 22.f4 can be met by Bronstein,D
♗h3! and the black queen is saved. Pein Beverwijk (Game 30) 1963
inverts the move order, but runs into [Steve Giddins]
something else. ]
21...♗g4?! QUESTION: What is the point of 1.c4 e6 2.♘c3 ♗b4 Diagram
this?
[ ANSWER: The threat was 21...-- 22.♗g2 (Diagram 31)
♕h2 23.♘f3 , trapping the queen, so
Bronstein's last stops that. ]
[ But objectively, 21...♘f6! is simpler and
stronger, when 22.♗g2 ♘g4 23.♕g3 ♕h2
is winning for Black. Bronstein was no
doubt seduced by the beautiful
combinative finish which occurs in the
game, but strictly speaking, his move
should have let White off the hook. ]

Bronstein − Move by Move

 Bronstein − 118
Move by Move
XHGFEDCBAY
1RsNLmKQvL-tR![
2zPPzPPzP-zPP" 31
3-+-+-sN-+#
4+-+-+Pvl-$
5-+-+-+-+%
6+-+p+-+-&
7pzpp+pzppzp'
8trn+kwqlsnr(
xhgfedcbay
QUESTION: I haven't seen that one before!
ANSWER: No, it is a surprisingly unusual
move, but was a favourite of Bronstein's.
Annotating the present game (in 'Shakhmaty
v SSSR', 4/1963), he wrote that he had
played "tens of games" with this line and
won almost all. The Mega Database contains
10 such games, with the splendid score of +6,
=4, so Bronstein had some grounds for his
claim that he regarded it as "my defence".
QUESTION: So what is so great about it?
ANSWER: What attracted Bronstein was its
flexibility and the lack of theory. As he
pointed out, Black still retains the chance to
play the Dutch, Nimzo-Indian, Ragozin,
QGD or some sort of irregular hybrid. He
particularly liked to follow up by putting his
knight on e7, rather than f6, and heading for
an unexplored formation, where both players
would have to think for themselves.
3.♕b3
[ One of a plethora of reasonable moves,
which include 3.♘f3 , ]
[ 3.g3 , ]
[ 3.e4 , ]
[ and even 3.a3 . Langeweg's choice
Bronstein − Move by Move
prevents his pawns being doubled. ]
3...♘c6 4.♘f3 ♘f6 5.g3
[ B r o n s t e i n c o m m e n t e d t h a t 5.a3
was the natural follow-up to White's 3rd
move, and I agree. After ♗xc3 6.♕xc3
White must objectively be a little better,
thanks to his two bishops and the fact that
he has not committed his d-pawn to d4 yet,
and thus can still opt fo r d3, keepin g
control of the e4-square. ]
5...d5 6.♗g2 0-0 7.0-0 d4 8.♘b1
QUESTION: So what is going on here?
ANSWER: Well, I do not really like the way
White has handled the opening. His knight
has been driven back to its starting square,
he has ceded central space, and his queen
excursion to b3 has proved a waste of time.
Bronstein, however, suggests that "Black
n e e d s t o s h o w ma x i m u m a t t e n t i o n a n d
caution", as White plans to advance on both
flanks, playing around the d4-pawn.
QUESTION: What about undermining the
pawn by e2-e3?
ANSWER: That is the other feature of the
position, and Bronstein comments that he
delved deeply into the position at this point,
"in order to find a sequence of moves in
which e3 would not be favourable for
White".
a5 QUESTION: So what is this for?
ANSWER: Black takes prophylactic
measures against a later a3 and b4 by White.
9.d3 e5 What we have now is a typical
Sch mid Benon i struct ure, with colours
reversed. This would typically be reached
after a sequence such as 1 d4 Nf6 2 Nf3 c5 3
d5 d6 4 Nc3 g6 5 e4 Bg7. Basically, Black
here will play for the breakthrough ...e5-e4,
whilst White must fight against this, and try
to get counterplay on the queenside or in the
centre by a later e3.
10.♘a3 A typical development plan in such
 Bronstein − 119
Move by Move
structures. The white knight is better on c2
than on d2 in such positions, because from c2,
it supports b4 and also e3. However, here
White has a less favourable version than
usual, because his queen is misplaced and
his queen's knight is already making its third
move.
♘d7 Again, a typical manoeuvre for such
positions. The king's knight's standard
position in these structures is c5, freeing the
way for ...f5 and ...e4 later on.
11.♕d1 Acknowledging the fact that his
early Qb3 has failed to justify itself.
♗e7 QUESTION: But Black also has to lose
a tempo!
ANSWER: True. His bishop needs to retreat,
before the knight can occupy c5.
12.♘c2 ♘c5 13.b3
[ Now we see the effect of Black's
prophylactic knight march to c5. He has
managed to prevent the break e3, since
13.e3 dxe3 leaves the d3-pawn hanging. ]
[ White therefore concentrates on preparing
b4, but even this he has to do the slow way,
since the immediate 13.a3 runs into a4
, either now or later. ]
13...♗g4 QUESTION: Why does the bishop
come here? Isn't it just going to get kicked
by h3?
ANSWER: It is another typical manoeuvre in
such positions. Black is quite happy to see
White play h3 and even g4, as this weakens
his position severely. Meanwhile, if the
bishop is not chased away, it eyes the e2-
pawn rather uncomfortably, and there is also
the possibility of ...Qd7 and a later ...Bh3,
trying to exchange off White's important
king's bishop.
14.♖b1 f5 Black seizes more space in the
centre and White's position is becoming
worse with every move. The knight on c5 is
key, as it stops a central response with e3.
Bronstein − Move by Move
15.♗a3 Diagram
XHGFEDCBAY
1-mKR+Q+R+!
2zPLzPP+N+P" 32
3-zPN+P+PvL#
4+l+-zpP+-$
5-+pzp-sn-zp%
6+-+-+n+-&
7pzp-vl-zpp+'
8+ktr-wq-+r({
xhgfedcbay
e4?! Bronstein plays very directly.
[ He admits that he initially wanted to play
15...♘e6 , which is very strong, ]
[ whilst the computer likes 15...♕d6 . ]
QUESTION: But I thought ...e4 was Black's
main idea anyway, so why not play it while
he can?
ANSWER: Well, the argument is that White
cannot effectively prevent it anyway, so
Black can afford to prepare it further, on the
basis that when it does eventually come, it
will be even stronger. But Bronstein's move
is also good. Black simply stands much better,
after any reasonable move.
QUESTION: So how come White stands so
badly?
ANSWER: His treatment of the opening was
just not good. The Schmid Benoni does not
enjoy a great reputation at the best of times,
and here, White has reached an inferior
version, where he lost time with his queen's
knight and queen.
16.♗xc5? This further mistake leaves White
effectively busted.
QUESTION: So what should he do?
 Bronstein − 120
Move by Move
[ ANSWER: He had to take Black on with
16.dxe4! .
A) Bronstein rejected this because of
16...d3 , when he gave the long forcing
variation 17.♘e3 ( 17.exd3? fxe4
18.dxe4 ♕xd1 wins a piece for Black)
17...dxe2 18.♕xe2 fxe4 19.♘xg4 exf3
20.♗xf3 ♘d4 21.♗d5+ ♕xd5 22.♕xe7
♕c6 . The line continues
A1) however, there are a couple of
things wrong with this variation; most
importantly, here White has the vital
improvement 23.f4! when he is holding
t h e b a l a n c e : f o r e x a m p l e , ♘f3+
24.♖xf3 ♕xf3 25.♕xc5 ♕xg4
26.♕xc7 and White has two pawns for
the exchange, plus a strong bishop - the
computer shows that White has equal
c h a n c e s h e r e , a s s h o w n b y ♖fd8
( 26...♕e2 27.♗xf8 ♕e3+ 28.♔g2
♕e2+ is an immediate draw) 27.♕e5
♖e8 28.♕d5+ ♔h8 29.♖d1
and neither side can achieve anything,
thanks to the weakness of both kings;
A2) 23.♗xc5 ♘f3+ 24.♔h1 ♖ae8
"and the queen is trapped", although
after ( instead, 24...♘h4+ 25.f3 ♖xf3
is immediately devastating) 25.♕xf8+
♖xf8 26.♗xf8 ♘d2+ 27.♔g1
Black still has work to do.;
B) In view of this, Black should meet 16
d x e 4 ! w i t h 16...♘xe4 , but he is not so
m u c h b e t t e r a f t e r 17.♗xe7 ♕xe7
18.♘fxd4 ♖ad8 ( 18...♘xd4 19.♕xd4
♗xe2 20.♖fc1 is equal ) 19.f3 ♘xg3
20.hxg3 ♘xd4 21.♘xd4 ♕e3+ 22.♖f2
♗h5 23.♕d3 ♕xd3 24.exd3 ♖xd4
. Black has an edge in this endgame, but
it not a huge amount.;
C) 16...-- ]
16...♗xc5 Now d4 is securely defended and
Black's positional advantage becomes
Bronstein − Move by Move
crushing.
17.♘fe1 ♕e7 18.f3 The only defence to the
threat of 18...exd3, winning e2.
♗h5? The start of a typically imaginative
id e a b y B r o n s t ei n , b u t o n ce a g a i n , t h e
ruthless silicon circuits of the computer
suggest it was not objectively strongest.
[ B l a c k s h o u l d p r e f e r s i m p l y 18...exf3
when he is much better after any recapture,
say, 19.exf3 ♗h5 and Black follows up
with ...f4, ...Rae8, etc. White's pieces are
so clumsily placed that he cannot organize
a defence. His two knights have only the
useless a1-square between them, for
example, and he cannot do anything to
challenge the open e-file. ]
19.♕c1? QUESTION: Can't he take on e4?
[ ANSWER: Indeed, he absolutely had to.
Bronstein's extraordinary idea was 19.fxe4
fxe4 20.♗xe4 and now the queen sacrifice
♕xe4!? .
QUESTION: Wow! Is that really good?
ANSWER: Alas, it appears not. Bronstein
gives 21.♖xf8+! ( 21.dxe4? d3+ 22.♔g2
♖xf1 23.♔xf1 dxe2+ 24.♕xe2 ♖f8+
25.♘f3 ♗xf3 wins, as shown by
Bronstein ) 21...♖xf8 22.dxe4 d3+ 23.♔h1
♖f1+ 24.♔g2 ♖g1+ 25.♔h3 ♗xe2
26.♕d2 ♘e5 27.♕g5 ♗f1+ 28.♔h4 ,
A) c o n t i n u i n g 28...♗e7 "and White's
position is crushed".
QUESTION: So what is wrong with that?
It looks convincing enough to me!
ANSWER: Unfortunately, after 29.♕xe7
♘g6+ 30.♔g5 ♘xe7 31.♘xd3 ♗xd3
32.♖xg1 ♗xc2 33.♖c1 White is actually
winning: for example, ♗d3 ( 33...♗xe4
34.♖e1 wins ) 34.♖d1 ♗c2 35.♖d2 ♗b1
36.♔f4 and the bishop is lost after Rb2.;
B) QUESTION: So are there no
improvements for Black in this line?
ANSWER: Yes, in fact the sacrifice is
 Bronstein − 121
Move by Move

good enough for a draw. Instead of 28...

B e 7 ? , B l a c k c a n d r a w b y 28...dxc2!
XHGFEDCBAY
29.♕xe5 ( 29.♘xc2? ♘f3+ wins ) 1-mKRsN-wQR+!
29...cxb1♕ 30.♕d5+ ♔f8 31.♕xc5+ 2zPL+P+N+-" 33
♔f7 32.♕f5+ and it's perpetual check.;
C) 28...-- ] 3-zPPzpP+PzP#
[ QUESTION: A pity! The computer always
seems to find mistakes in these analyses!
4+-+-zpP+-$
ANSWER: That is true, although there are 5l+p+-vl-zp%
exceptions. Back at move 23, we had this
position: 19.fxe4 fxe4 20.♗xe4 ♕xe4 6+-+-+n+-&
21.♖xf8+ ♖xf8 22.dxe4 d3+ 7pzp-wq-zpp+'
. Here, instead of
A) the move 23.♔g2 is another try, when 8+ktr-+-+r({
Bronstein's analysis was very accurate:
♖f2+ 24.♔h3 ♘e5 25.exd3
xhgfedcbay
( Bronstein also mentioned the stronger ♖a6! The rook finds the shortest path to the
move 25.g4! and then ♗xg4+ 26.♔g3 h-file.
♗xe2! 27.♕d2 ♘f3 28.♘xf3 ♖xf3+ 21.b4 White battles desperately for some air,
29.♔g2 ♖f2+ 30.♔g3 ♖f3+ but it is all hopeless.
with perpetual; the computer cannot axb4 22.axb4 ♘xb4 23.♘xb4 ♗xb4
improve on this analysis at all) 25...♗xd1 24.♘c2 Diagram
26.♖xd1 h5 "and Black's attack is very
dangerous". The computer agrees and
XHGFEDCBAY
assesses this as clearly better for Black: 1-mKR+-wQR+!
for example, 27.g4 (forced) ♘xg4 28.d4 2zPL+P+N+-" 34
♖xh2+ 29.♔g3 ♗e7 and Black is
better.; 3-zPPzpP+-+#
B) 23.♔h1 , ]
19...e3 Now, on the other hand, it really is all
4+-+-zpPvl-$
over. Black's crushing central pawn phalanx 5l+p+-+-+%
slices White's position in half, with his
queen, rook and knight stuck in the wrong 6+-+-+-+r&
half. While they are desperately trying to 7pzp-wq-zpp+'
free themselves and get back into the action,
Black will take aim at the enemy king, via 8+ktr-+-+-({
the weak dark squares, and that will be that.
20.a3 Diagram
xhgfedcbay
♗c3! QUESTION: Why is the bishop good
here? It actually looks a little cut off.
ANSWER: The important point is to keep
White's position cut in half. It prevents the

Bronstein − Move by Move

 Bronstein − 122
Move by Move

white queen coming to the kingside via e1.

25.♖xb7 f4 26.g4 ♗e8 27.♕b1 ♖h6
No w the final, fatal infiltration on the
kingside dark squares begins.
28.♘b4 ♕h4 29.h3 ♕g3 30.♔h1 ♖xh3+
[ I t i s mat e a f t e r 30...♖xh3+ 31.♗xh3
♕xh3+ 32.♔g1 ♕g3+ 33.♔h1 ♖f6 . ]
0-1

Bronstein − Move by Move